* Rename agnes-metadata.json to marketplace-metadata.json
Curated marketplace enrichment file (.claude-plugin/agnes-metadata.json)
becomes marketplace-metadata.json. Clean cut, no fallback — curators of
upstream marketplace repos must rename the file on their side.
Python API renames mirror the file rename: read_agnes_metadata →
read_marketplace_metadata, AGNES_METADATA_REL → MARKETPLACE_METADATA_REL,
AGNES_METADATA_MAX_BYTES → MARKETPLACE_METADATA_MAX_BYTES. Synth Claude
Code marketplace strip rule (.agnes/** + the metadata file) follows the
new filename.
* Marketplace detail polish: window cover + 715:310 aspect + helper alignment
- Plugin & item (skill/agent) detail hero: 160x160 square cover replaced
with a macOS-style window frame (3 traffic-light dots + titlebar label
showing the entity name). Body is constrained to 715:310 so curator-
uploaded covers no longer crop to a square. Window is 380px wide; meta
column and absolutely-positioned top-right install/remove actions stay
put. Fallback when no cover_photo_url (translucent gradient + PL/SK/AG
initials) is unchanged, just inside the window body.
- Inner skill/agent cards in the plugin detail's Internal structure
section adopt the same 715:310 aspect (was fixed 78px tall). No window
chrome on inner cards — just the matching proportions so covers read
consistently across hero, grid tiles, and listing cards.
- Curated nested item helper text ("This skill is part of ... — add the
bundle to your stack to use it") now stacks UNDER the "Open parent
plugin" button instead of being a side-by-side flex sibling in the
actions-row. Added align-self: flex-end so the 260px helper box
anchors at the right edge of the 300px actions column, matching the
button's right edge.
* Marketplace My tab: surface the same category + type filters as Flea
- Frontend: mp-cat-row and mp-type-row now show on tab=my (previously
hidden — type was flea-only, category was flea/curated-only). Curated
browse stays plugin-only and continues to hide the type pills.
fetchOne() sends the `type` param for tab=my too, so the items
endpoint's existing my-branch filter actually receives it.
- Backend categories endpoint, tab=my branch: when the type filter is
set to skill/agent, skip counting curated subscriptions. Curated
plugins are always type='plugin', so they wouldn't survive the items
endpoint's type filter; including them in the category counts made
the pill numbers overstate what users could actually see in the
grid. type=None or type='plugin' keeps the previous behaviour.
- CHANGELOG entry under [Unreleased].
* Marketplace plugin detail: render rich content from marketplace-metadata.json
Adds five optional plugin-level fields to marketplace-metadata.json and
renders them on the curated plugin detail page + listing card:
* display_name — friendly h1 / listing-card name / mac-window titlebar
label (overrides the technical plugin id)
* tagline — punchy 1-line value prop for the hero subtitle and the
listing card description (replacing the verbose marketplace.json
description on cards)
* description — multi-paragraph markdown body, server-side rendered
through markdown-it-py and sanitized through nh3 with a
description-scoped allowlist (no iframes / no raw HTML / no
javascript: links). Powers the "What it does" panel.
* use_cases[] — {title, description, prompt} entries that render as a
3-column "When to use it" card grid; each card shows the literal
prompt as a code chip so users can copy-paste into Claude Code.
* sample_interaction — {user, assistant} dialog rendered in a Claude
Code-style dark Catppuccin Mocha transcript panel: monospace user
row with a green ">" prompt indicator + sans-serif assistant body
with markdown formatting (peach bold, yellow italic, pink inline
code, mantle-dark fenced code blocks).
All five fields are optional; UI sections only render when populated,
so plugins without enrichment look identical to before. Fields are
read on-demand from the working tree (cached by mtime per marketplace
slug) so curator edits land at the next request without waiting for
a sync cycle — same pattern as the existing inner-skill/agent
enrichment path. No DB schema bump.
Skill / agent rich-content rendering is deferred to a later phase
(needs a source-of-truth decision: extend plugin.yml? LLM-generate
from SKILL.md / agent.md?). The schema accepts the same fields at
skill/agent level today for forward compatibility but the UI ignores
them for now.
Also: stripped a stale `background-color: var(--bg)` from the global
`code` rule in style.css (was making inline code visually disappear
on the page background).
* Skill / agent detail: render rich content from marketplace-metadata.json
Brings the skill/agent detail pages to parity with the plugin detail
page. Same rich-content schema (display_name, tagline, description as
markdown, use_cases[], sample_interaction) plus two per-item additions:
* invocation — curator-provided literal command string. When set,
overrides the computed "<manifest_name>:<inner_name>" chip and
cleanly supports both "/" skill prefix and "@" agent prefix (the
hardcoded "/" in the chip markup is hidden when the curator provides
the invocation, so /grpn-eng:query <q> and @grpn-eng:cto-architect
both render correctly).
* when_to_use — markdown disambiguation block ("Use this for X. For
similar Y, see /other-skill") rendered into a new "When to use this"
panel below the Example section.
Skill / agent category is now per-item overridable in
marketplace-metadata.json. When absent, the API keeps the parent
plugin's category as the badge so existing items don't lose their
category until curators opt in to per-item categorization.
The new "Example" Q&A panel uses the same Claude Code-style dark
Catppuccin Mocha transcript treatment as the plugin detail —
monospace user row with a green ">" prompt indicator + sans-serif
assistant body with markdown formatting.
All new fields are optional and read on-demand from the working tree.
Skills / agents whose marketplace-metadata.json doesn't carry rich
content render exactly the same way they did before (frontmatter
description + computed slash command + cover from existing v32
enrichment). No DB schema bump.
* Fix TypeError in skill / agent detail when curator sets per-item category
`curated_skill_detail` and `curated_agent_detail` were passing both
`**parent` (from `_curated_inner_parent_fields`, which returns the
parent plugin's category as a fallback) and `**enrichment` (from
`_curated_inner_enrichment`, which returns the per-item category
override when the curator set one) into `InnerDetailResponse(...)`.
Python function-call kwargs unpacking with overlapping keys raises
`TypeError: got multiple values for keyword argument 'category'`
— it doesn't merge like a literal dict does. The bug only surfaced
when the marketplace-metadata.json carried a `category` field at
skill / agent level (curator opting into per-item categorization);
items without that override hit the endpoint cleanly because only
parent provided the key.
Fix: build `merged = {**parent, **enrichment}` first (literal-dict
syntax DOES merge, with the right-hand-side winning) and unpack the
merged dict. Curator override still wins via the merge order, and
the same pattern is future-proof for any other field that lands in
both layers later.
Plus a regression test in test_marketplace_metadata.py asserting
that the inner-resolver carries `category` for downstream merging.
* Marketplace detail: tolerate partial curator JSON
Server constructed UseCase / SampleInteraction via raw dict indexing
(uc["title"], sample["assistant"]), so a curator commit missing any
required Pydantic field crashed the whole plugin / skill / agent detail
endpoint with a 500. Route both constructions through _safe_use_case /
_safe_sample_interaction helpers — partial input silently drops the
malformed card / section instead of breaking the page.
Regression test in test_marketplace_api.py covers the three shapes:
use_case missing a key, use_case with an empty string, and
sample_interaction with only user (no assistant). Sibling rich fields
still render.
* Address PR-251 review (must-fixes + S2/S3 polish) + release-cut 0.50.0
Five must-fixes from the review pass (3 from @cvrysanek's two-stage
review, 2 from my independent pass), plus the 0.50.0 release-cut as the
last commit on this PR per CLAUDE.md (CLAUDE.md "Release-cut belongs
to the PR" rule added in v0.49.1).
Must-fixes
----------
1. Cache eviction: bounded LRU instead of per-marketplace predicate.
The previous predicate (`k[0] == marketplace_id and k[1] != mtime_ns`)
only swept stale entries for the CURRENT marketplace; with N>100
distinct marketplaces each holding one mtime key, the cap silently
failed and memory grew linearly. Replaced with OrderedDict-backed
bounded LRU at cap=256, drop oldest insert on overflow.
Cache stress test pinned in test_marketplace_metadata.py.
2. Render CPU cap: per-field byte cap on description / when_to_use /
sample_interaction.assistant via MARKETPLACE_METADATA_FIELD_MAX_BYTES
(= 64 KiB). Without this, a 1 MiB curator markdown body × QPS =
curator-controlled CPU burn through pure-Python markdown-it-py.
Truncation respects UTF-8 boundaries and logs a warning so the
curator sees the cap fire on the next sync. Test for cap +
UTF-8-boundary preservation.
3. Inner-detail bypassed the metadata cache. _curated_inner_enrichment,
_curated_inner_cover, and curated_detail all called
read_marketplace_metadata directly, defeating the mtime cache the
plugin listing already shared. Routed all three through
_read_metadata_cached so skill/agent detail hits are O(1) re-parses
per marketplace per mtime instead of O(QPS).
4. Truthy-vs-presence trap in plugin/inner enrichment merge. API-layer
writers used `if resolved.get(k):` which silently dropped any
future falsy-but-valid resolver field (bool featured=False, int
priority=0, str category=''). Switched to presence check
(`if k in resolved`) so the resolver is the authority on field
presence; `{**parent, **enrichment}` merge respects whatever the
resolver decided to ship.
5. Vendor-agnostic OSS cleanup. Removed operator-specific token
references (/grpn-eng:, @grpn-eng:, .foundryai/) from
src/marketplace_metadata.py docstring, app/web/templates/
marketplace_item_detail.html JS comment, docs/curated-marketplace-
format.md, and tests/test_marketplace_metadata.py fixtures. Replaced
with generic /my-plugin:tool / @my-agent:role / .example/ placeholders.
CHANGELOG
---------
- New "### Fixed (PR #251 follow-ups)" section documenting all 4
code-side must-fixes
- New "### Internal" section noting the vendor cleanup + new tests
- BREAKING bullet for the file rename now covers operator-side
migration: running instances see plugin enrichment disappear from
the UI until upstream curator renames + nightly sync overwrites the
working tree; POST /api/marketplaces/{id}/sync forces refresh sooner
- Stripped /grpn-eng: leaks from the existing skill/agent rich-content
bullet
Tests
-----
128 targeted tests pass (test_marketplace_metadata, test_marketplace_api,
test_marketplace, test_markdown_render, test_marketplace_synth_strip,
test_marketplace_filter). New tests added:
- 6 XSS regression tests on render_safe (javascript:/data:/vbscript:
schemes via autolink, reference link, and mixed-case + positive
http/https/mailto + noopener noreferrer rel)
- 3 byte-cap tests (truncation + UTF-8 boundary + under-cap pass-through)
- 1 cache eviction stress test (>256 marketplaces -> bounded at cap)
- 1 truthy-vs-presence resolver-contract test
Release-cut
-----------
- pyproject.toml 0.49.1 -> 0.50.0 (minor; BREAKING file rename per
pre-1.0 CHANGELOG note: "breaking changes called out under Changed
or Removed with the BREAKING marker")
- CHANGELOG [Unreleased] -> [0.50.0] - 2026-05-12, new empty
[Unreleased] on top.
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
Co-authored-by: ZdenekSrotyr <zdenek.srotyr@keboola.com>
Three behavioural improvements driven by the sub-agent end-to-end test
findings, plus scheduler tweaks to prevent the post-deploy contention
burst we measured.
CATALOG (catalog-side bugs the test agents tripped on):
- new entity_type field per remote row (BASE TABLE / VIEW /
MATERIALIZED VIEW). For views, rows + size_bytes return null
instead of the misleading 0 that __TABLES__ reports.
- where_examples now validates against the table's actual schema
(cached known_columns from refresh). The pre-fix behavior
blindly advertised `country_code = 'CZ'` on tables with no
country_code column — the sub-agent tests reliably hit this on
unit_economics.
- new known_columns + entity_type columns on bq_metadata_cache;
populated by bq_metadata_refresh.refresh_one from the same
fetch_bq_columns_full call (no extra BQ roundtrip) plus a
cheap INFORMATION_SCHEMA.TABLES lookup for table_type.
QUERY COST-GUARD:
- remote_scan_too_large suggestion now names views explicitly:
`Target(s) <ids> are VIEW or MATERIALIZED VIEW. BigQuery does
not push LIMIT into the view body — SELECT * FROM <view>
LIMIT 1 still runs the full underlying scan.` Programmatic
consumers get a new view_targets field on the error detail.
SCHEDULER HYGIENE (the post-deploy 1-minute window where
concurrent parquet downloads dropped to ~1 MB/s):
- SCHEDULER_STARTUP_GRACE_SECONDS (default 60) holds the first
tick so the burst doesn't overlap cache_warmup writes.
- SCHEDULER_BQ_METADATA_INITIAL_OFFSET_MAX_SECONDS (default 900)
randomises bq-metadata-refresh's first-fire offset.
TESTS:
- test_bq_metadata_cache_repo: entity_type + known_columns round-trip
- test_v2_catalog_remote_metadata: where_examples validation, views
return null rows/size_bytes, cold rows have empty examples
- test_api_query_guardrail: VIEW-aware suggestion text + view_targets
- test_connectors_bigquery_metadata: entity_type lookup mock + new
fields in TableMetadata expectations
- test_scheduler_sidecar: grace + jitter env-var resolution
* fix(cta): fall back to textarea+execCommand when Clipboard API rejects
The "Setup a new Claude Code" CTA fetches /auth/tokens, parses the JSON
response, renders the setup script, THEN calls
`navigator.clipboard.writeText()`. Modern browsers (Safari, Firefox, and
Chrome on stricter configurations) reject `writeText` with
NotAllowedError when transient user activation has been consumed by an
intervening `await` — which is exactly the case here. Users perceived
this as "the browser blocked the copy" and got the manual-paste fallback
modal even though the textarea + `document.execCommand('copy')` path
WOULD have worked synchronously without needing fresh user activation.
`copyToClipboard` now:
- prefers the modern Clipboard API (unchanged for the happy path)
- on writeText rejection, falls back to `copyViaTextarea` instead of
surfacing the rejection to the caller's catch block.
`copyViaTextarea` is the previously-inline textarea fallback factored
out into a named helper, with two small hardening touches:
- `readonly` + `tabindex=-1` so the hidden textarea doesn't steal
focus or pop the virtual keyboard on mobile.
- explicit `setSelectionRange(0, text.length)` to belt-and-braces the
selection on iOS Safari (where `.select()` alone sometimes selects
zero chars on touch-focused textareas).
Only the CTA button needed this — the Step-1 install-command and the
connector-copy buttons all call `writeText` synchronously inside the
click handler (no awaits in between), so they keep their existing
user-gesture context and didn't hit the same rejection. No template
changes there.
* refactor(home): fold Atlassian MCP registration into connectors block
The standalone "Register the Atlassian MCP server" step (was step 6 in
the unified setup script) moves INTO the Atlassian connector's prompt
body so all Atlassian-related setup lives in one logical group. Same
intent that #247 carried for connectors, applied one level deeper:
the hosted Remote MCP registration is part of "set up Atlassian", not
its own ungrouped step.
What changed:
- `app/web/connector_prompts.py` — the Atlassian prompt's step 5
replaces the speculative "Register the on-demand Atlassian MCP under
.claude/mcp/atlassian" line with the actual hosted Remote MCP
registration: `claude mcp add --transport sse atlassian
https://mcp.atlassian.com/v1/sse || true`. The `|| true` keeps re-runs
idempotent and the body explains the OAuth-on-first-use contract.
Both /home's Atlassian tile and the inlined setup-script Atlassian
sub-block emit this line — single source of truth holds.
- `app/web/setup_instructions.py` — `_mcp_servers_block` deleted; the
`mcp_servers` step is removed from `_step_numbers`; resolve_lines no
longer calls it.
- Renumbering: install (1), init (2), catalog (3), preflight (4),
marketplace (5), diagnose (6), connectors (7), confirm (8). Was:
6 = mcp_servers, 7 = diagnose, 8 = connectors, 9 = confirm.
- `tests/test_setup_instructions.py` — Confirm step 9→8, Connect 8→7,
diagnose 7→6, mcp_servers references dropped.
`test_step_numbering_with_connectors_step` now asserts
`"mcp_servers" not in steps`. Stray-Confirm assertion lists shift
by one position.
- `tests/test_setup_page_unified.py` + `tests/test_web_ui.py` — same
step-number shifts in the rendered /setup preview assertions.
The `claude mcp add` line is still the Atlassian Remote-MCP path that
the 2026-05-10 init-report Fix C added — only its position in the
flow changes. /home Atlassian tile copying continues to install the
MCP too (the prompt body the tile pastes contains the same line).
112 tests pass.
* feat(atlassian): operator-overrideable base URL via AGNES_ATLASSIAN_BASE_URL
Adds an env var / YAML key the operator (Terraform module, customer-VM
template, OSS instance.yaml) can set to bake the Atlassian Cloud site
root into the connector prompt — so end users don't have to guess /
paste their org's `https://<myorg>.atlassian.net`.
When set, the Atlassian connector prompt (rendered on both /home tile
and inlined into the setup-script step 7 Atlassian sub-block) replaces
step 1's "Ask me for my Atlassian Cloud site URL and email" with a
one-line note that the URL is already provisioned by the operator and
asks only for the email. Step 4's helper-script body has the
`BASE_URL='<the site URL I gave you>'` placeholder substituted with
the literal value. When unset (empty), the existing "ask the user"
flow remains — no regression for OSS instances.
Resolution + normalization in `get_atlassian_base_url()`:
- env `AGNES_ATLASSIAN_BASE_URL` > yaml `instance.atlassian.base_url` > ""
- strips trailing slash + trailing `/wiki` so the canonical value is
the bare site root. Matches the per-user helper script's
normalization at storage time (atlassian_prompt step 4 guard 2), so
the literal baked in by the operator stays consistent with what the
user's helper script would have computed from their input.
Plumbing:
- `app/instance_config.py`: new `get_atlassian_base_url()` resolver.
- `app/web/connector_prompts.py`:
- `atlassian_prompt(*, base_url: str = "")` — string-replace two
explicit placeholder phrases when base_url is truthy; otherwise
return the prompt unchanged.
- `all_connector_prompts(..., atlassian_base_url: str = "")` —
forwards the kwarg.
- `app/web/router.py` (`_build_context`): reads
`get_atlassian_base_url()` and passes it through to
`all_connector_prompts(...)` so both the /home tile context AND the
inlined-script `resolve_lines(...)` call use the same value.
- `src/welcome_template.py` (`compute_default_agent_prompt`): same
threading via the existing import-on-demand path.
Tests (`tests/test_home_route_resolution.py`):
- `get_atlassian_base_url` resolver: default empty, env override,
trailing-slash strip, trailing-`/wiki` strip.
- `atlassian_prompt(base_url=...)`: literal URL baked in, ask-step
removed, placeholder replaced, operator-baked-in copy appears.
- `atlassian_prompt(base_url="")`: existing ask-the-user flow
unchanged.
- `all_connector_prompts(atlassian_base_url=...)`: kwarg threads
through to the rendered atlassian prompt.
135 tests pass.
* feat(asana): register hosted Asana Remote MCP in connector prompt
The Asana connector prompt only stored a PAT in the OS keychain + ran
a curl verify against /api/1.0/users/me. That set Claude Code up for
direct `curl` calls but didn't actually wire Asana into Claude's tool
list — so the user couldn't ask Claude to "find my open Asana tasks"
and have it work. Symmetric oversight to the Atlassian connector's
original speculative `.claude/mcp/atlassian` line that this branch
already replaced with `claude mcp add --transport sse atlassian
https://mcp.atlassian.com/v1/sse`.
Adds a new step 5 that registers Asana's hosted Remote MCP:
claude mcp add --transport http asana https://mcp.asana.com/mcp || true
This is the V2 endpoint (streamable HTTP transport, launched February
2026). The V1 SSE endpoint at https://mcp.asana.com/sse was deprecated
2026-05-11 (today) and must NOT be used — calling it out explicitly
in the prompt body so a future operator who finds an old reference
doesn't paste the dead URL. OAuth is handled by Claude Code at first
use, same model as the Atlassian MCP step.
The PAT stored in step 3 stays for direct `curl` calls (precheck +
ad-hoc scripts) — the MCP path uses its own OAuth grant, not the PAT.
Old step 5 (revoke instructions) renumbers to step 6 and adds the
`claude mcp remove asana` cleanup hint.
Same single-source-of-truth invariant holds: /home Asana tile + the
inlined Asana sub-block in the setup script (step 7 connectors) both
emit identical text from `asana_prompt()`.
71 tests pass.
* feat(asana): drive MCP OAuth login + end-to-end validation post-register
`claude mcp add --transport http asana ...` only registers the
server in Claude Code's local config — it does NOT trigger OAuth.
The browser tab opens the first time any `mcp__asana__*` tool gets
invoked. So the previous step 5 left a user looking at a "registered"
MCP that, in practice, hadn't authed yet and would fail on first
real use. Same blind spot Atlassian's prompt also has, but Asana was
the one called out in the latest review pass.
Adds a new step 6 between MCP registration (step 5) and the revoke
instructions (now step 7):
a. Tell the user verbatim what's about to happen — a low-impact
read through the MCP will pop the OAuth browser tab; sign in
with the same account whose PAT they stored in step 3 and
approve. Frames the OAuth as one-time so users don't wait
for it on every later call.
b. Drive an actual MCP read. Don't prescribe the exact tool name
because the Asana MCP's exposed surface (`mcp__asana__*`) is
versioned upstream and we don't want to pin to a name that
gets renamed. Instead: tell Claude to pick the lightest read
from its surfaced tool list (users-me / list-workspaces /
equivalent). Document the recovery path when Claude Code
times out waiting for the OAuth tool use: `claude mcp list`
to confirm registration before retrying.
c. Print a single one-line proof that combines wiring + auth:
"Asana MCP connected as <name> — <N> workspace(s) visible."
Explicit anti-echo callout for tokens, task content, comments.
On failure, surface the exact Claude-Code error and stop —
no silent pass.
d. Sanity-check that the MCP OAuth identity and the PAT identity
reference the same Asana account. Easy mistake to make when
the user has multiple Asana accounts — flag only on mismatch,
keep quiet when they match. Recovery: `claude mcp remove asana
&& claude logout asana` then redo step 5.
Step 7 (revoke) absorbs both the keychain delete + the
`claude logout asana` line so users have a single place to undo
everything.
43 tests pass.
* fix(init): clear stale CA env vars on Windows before any TLS handshake
Reported by the 2026-05-11 Windows test pass: after `agnes init` the
gws connector failed with `UnknownIssuer` TLS errors because
`SSL_CERT_FILE` and `REQUESTS_CA_BUNDLE` were still set in Windows
User scope pointing at `C:\Users\localadmin\.config\agnes\ca-bundle.pem`
— a file that did not exist on the test host. Past Agnes installs
(the setup-prompt trust block + older bootstrap helpers) write those
pointers when they materialize a combined Agnes-CA bundle; when the
bundle file later disappears (re-init on a new VM, machine swap, the
~/.agnes dir wiped), the pointers go stale and every native Windows
TLS handshake fails before Agnes itself runs. SSL_CERT_FILE in
particular REPLACES (not appends to) the trust store, so a stale
pointer is silently catastrophic.
`agnes init` now clears stale pointers in two layers before the first
server roundtrip:
1. Current-process env (os.environ) — what the immediately-following
`api_get` to /api/catalog/tables actually reads. Without this, init
itself blows up before it gets to step 2.
2. Windows User-scope env via PowerShell
`[Environment]::SetEnvironmentVariable(name, $null, 'User')` — what
every future shell + every native tool (gws, claude.exe, pip, uv)
inherits. The 2026-05-11 reporter expected this exact cleanup
("init was supposed to clear these but they persisted").
The cleanup is best-effort and conservative:
- Only deletes a var when its value points at a path that does NOT
exist on disk. Intentional operator config (e.g. SSL_CERT_FILE
pointing at a corp certifi bundle) stays put.
- PowerShell missing / restricted execution policy / WSL-without-pwsh:
swallowed silently. The current-process leg still runs, which
unblocks init even on hosts where the User-scope leg cannot fire.
Tests (`tests/test_init_ca_cleanup.py`, 6 cases):
- Stale pointers → removed from process env.
- Real-path pointers → preserved.
- Non-Windows hosts: PowerShell is not invoked.
- Windows hosts: PowerShell IS invoked with a script that checks
all three vars + uses Test-Path + SetEnvironmentVariable.
- PowerShell FileNotFoundError: cleanup swallows it, does not raise.
- `_is_windows_host()` reflects sys.platform.
* refactor(asana): MCP-first flow — drop PAT storage, precheck via `claude mcp list`
The Asana hosted MCP at https://mcp.asana.com/mcp authenticates via
OAuth (Claude Code holds the grant; browser tab pops on first tool
use). The earlier prompt walked the user through creating + keychain-
storing an Asana Personal Access Token AND registering the MCP — two
parallel auth surfaces for one connector. Once the MCP works, the PAT
has no consumer: the precheck/verify steps that used `curl
$BASE/api/1.0/users/me` are just redundant proof that Asana itself is
reachable, which the OAuth handshake already establishes.
Removed:
- Step 0 keychain probe + curl verify against /users/me with PAT.
- Step 1 open developer-console / create PAT.
- Step 2 click "+ New access token", warn shown-ONCE.
- Step 3 helper-script for keychain-storage (per-OS bodies: macOS
`security add-generic-password`, Linux `secret-tool store`, Windows
`cmdkey /generic`).
- Step 4 PAT-side `users/me` verify.
- Step 5's split that kept the PAT around for direct curl scripts.
- Step 6d's "MCP vs PAT identity sanity check" — there is no PAT
anymore, nothing to mismatch against.
New flow (3 steps total):
- Step 0 precheck: `claude mcp list | grep ^asana` — if found, the
server is registered AND Claude Code is holding its OAuth grant
(otherwise prior failure would have removed it); print
"Asana MCP already registered — skipping setup" and stop. Tells the
user the explicit reset command (`claude mcp remove asana && claude
logout asana`) so a re-register stays one paste.
- Step 1: `claude mcp add --transport http asana
https://mcp.asana.com/mcp` — no `|| true` because step 0 should have
caught the "already exists" case. Step explains the V2-vs-V1
endpoint distinction (V1 SSE deprecated 2026-05-11) and the
abort-clean recovery if the precheck somehow missed the existing
server.
- Step 2: same OAuth + low-impact-read validation pattern as before.
- Step 3: revoke instructions (mcp remove + logout + Asana-side app
revoke at app.asana.com/Settings → Apps).
Both surfaces (the /home Asana tile and the inlined Asana sub-block
in the setup script's step 7) emit the new text from the same
asana_prompt() — single-source-of-truth invariant intact.
77 tests pass.
- DuckDB 1.5.1 regressed: rejected `ALTER TABLE … ADD COLUMN IF NOT EXISTS`
with `Cannot alter entry … because there are entries that depend on it`
when the target was FK-referenced from another table. Hit on `internal_roles`
(v8→v9) and `user_groups` (v11→v12) during migration replay. 1.5.2 fixes it.
CI already runs 1.5.2; this pins the same floor for local devs.
- tests/test_cli_binary_rename now skips with an actionable message instead
of failing when the local venv has no `agnes` on PATH (fresh checkout) or
has a stale shim from a prior editable install whose `cli` layout shifted.
CI installs fresh and still asserts the real contract.
Since 0.47.0 GET /api/v2/catalog enriched each remote BigQuery row by
fetching INFORMATION_SCHEMA.TABLE_STORAGE + COLUMNS through the DuckDB
BigQuery extension *inside the request*. On cold caches that fanned out
to O(N) sequential BQ jobs-API roundtrips — easily 90 s+ on partitioned
/ view-backed tables — and reliably blew the CLI's 30 s httpx
ReadTimeout. Reproduced with py-spy: three AnyIO worker threads stuck
inside connectors/bigquery/metadata._fetch_via_legacy_tables.
Refactor: enrichment is read exclusively from a new persistent
bq_metadata_cache DuckDB table (schema v40), populated by a scheduler-
driven refresh job at SCHEDULER_BQ_METADATA_REFRESH_INTERVAL (default
4 h). Cold catalog response on a fresh container is now tens of
milliseconds with metadata_freshness=never_fetched for unwarmed rows.
New surface:
- POST /api/admin/run-bq-metadata-refresh (scheduler-driven, full)
- POST /api/v2/metadata-cache/refresh?table=<id> (admin, single)
- GET /api/v2/metadata-cache/status (auth, non-admin)
- metadata_freshness field per catalog row
Removed (internal API): v2_catalog._size_hint_for_row,
_resolve_remote_metadata, _metadata_provider_for,
_build_metadata_request, _materialized_size_hint, in-memory
_metadata_cache. Response shape unchanged for external consumers.
991 tests passing; 2 pre-existing failures (test_db v3→v4 ladder,
test_cli_binary_rename) unrelated to this change.
* Make /home install-hero links readable against blue background
The Claude license-options link added in the previous commit inherited
the default `<a>` style (`var(--hp-primary)` blue), which renders as
blue-on-blue and is unreadable inside the blue install-hero. Add a
scoped `.install-hero a` rule that uses white with an underline
(matching the existing lead-paragraph contrast pattern) so any link
nested in the hero stays legible.
* Reorder /home install flow: auto-mode is now Step 2, Agnes install becomes Step 3
Step 3 (was Step 2) pastes a ~20-command bash bootstrap into a fresh
Claude Code session. Without auto-mode enabled first, each Bash/edit
command needs a manual approve click — bad UX for first-time users.
Move auto-mode from the outside-hero `<details>` reference block into
the install-hero as a real Step 2, between "install Claude Code" and
"install Agnes". Content is the persistent `acceptEdits` snippet
(write to ~/.claude/settings.json) plus a one-liner pointing at
Shift+Tab for users who are already inside a running Claude Code
session. YOLO mode for full Bash auto-approve stays on
/setup-advanced behind the existing link.
The outside-hero `setup-collapsible[data-section="step3"]` block is
dropped — auto-mode is no longer reference content, it's a real
install step, and duplicating it would just diverge over time.
Onboarded users no longer see the auto-mode block at all (consistent
with Steps 1 + 3 also hiding post-onboarding).
Completion banner copy updated: "Step 1, 2 & 3 done — Claude Code
installed, auto-mode set, Agnes ready". Dashboard CTA partial and
other templates don't reference step numbers for this flow, so no
adaptation needed there.
* Simplify /home Step 2 to Shift+Tab only — drop the JSON snippet
Operator pointed out two issues with the prior Step 2:
1. The settings.json snippet is redundant. Claude Code's first
Shift+Tab cycle to auto-accept mode already prompts the user
whether to persist it as default — Claude writes the config
itself, no manual file edit needed.
2. The snippet only showed the POSIX path `~/.claude/settings.json`,
which doesn't translate to native Windows.
Replace the snippet + copy button with a plain Shift+Tab instruction,
explicitly call out the first-time "make this the default?" prompt,
and note that Claude handles the config write itself — same flow on
macOS / Linux / WSL / Windows. Adds a fallback line for users who
already closed the post-OAuth session.
* Tighten /home Step 2 install-note to two paragraphs
Operator: drop the 'Claude writes the setting itself, so this works
the same on macOS / Linux / WSL / Windows...' line plus the
'auto-approves file edits going forward; Bash commands stay gated
— that's the safe default' line. Both were filler — the make-default
prompt already implies persistence, and gated Bash is the obvious
default users won't be surprised by.
Result: paragraph 1 carries Shift+Tab + first-time make-default
say-yes + closed-session fallback in one breath; paragraph 2 keeps
the verbatim YOLO link. Same affordances, less vertical space.
* Capture session paths via SessionStart hook + lock parallel pushes
Replace the encoding-based scan of ~/.claude/projects/<encoded-cwd>/ with
a queue file populated by a new `agnes capture-session` SessionStart hook.
The hook reads the documented `transcript_path` field from Claude Code's
hook stdin JSON, sidestepping the cwd-to-folder encoding (which is an
internal implementation detail and varies by Claude Code version).
- New `agnes capture-session` subcommand appends transcript_path to
<workspace>/.claude/agnes-sessions.txt. Silent on all malformed input
so a hook chain failure doesn't clutter Claude Code startup.
- `agnes push` now consumes the queue: atomic snapshot rename guards
against hooks writing during the push window, successful uploads land
in agnes-sessions-uploaded.txt (TSV: timestamp + path), failed paths
are requeued.
- Cross-platform single-instance lock via the filelock package (fcntl
on POSIX, msvcrt on Windows). Concurrent SessionEnd hooks — common
when the user closes several sessions at once — silent-exit on the
losing side instead of all racing the upload.
- Recovery: pre-existing snapshot files from a crashed push are picked
up and processed before the live queue.
- The SessionStart `agnes push` self-heal entry is dropped — it became
redundant once the queue persists across runs (orphans from headless /
crashed sessions ship out on the next interactive SessionEnd push).
Existing workspaces auto-migrate via the marker-based replace logic.
- Legacy encoding scan stays available behind `--legacy-scan` for one-
off backfills of sessions predating the queue.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Add /agnes-private + statusLine indicator for private sessions
Users handling sensitive data inside Claude Code can now opt a session
out of the Agnes upload pipeline, either proactively (right after session
start) or reactively (mid-session). The `/agnes-private` slash command
runs `agnes mark-private` deterministically via `!`-prefix direct bash —
no AI in the loop. A workspace-installed statusLine surfaces a
`🔒 agnes-private` indicator in Claude Code's status bar so the user
sees the state at a glance.
Authoritative source of "do not upload" is a separate file
`<workspace>/.claude/agnes-sessions-private.txt` (one session_id per
line). Both `capture-session` (queue writer) and `push` (queue reader)
consult the list. This makes the slash-command / SessionStart-hook race
impossible by construction: whichever runs first, the session is correctly
filtered out.
- `agnes mark-private` reads `CLAUDE_CODE_SESSION_ID` from env (set by
Claude Code in every bash subprocess it spawns — stable documented API)
and appends to the private list.
- `agnes statusline` reads the session JSON Claude Code pipes on stdin,
checks the private list, and emits the indicator or nothing. Optimized
for the high call frequency of statusLine renders.
- `capture-session` extracts session_id from hook stdin and skips queue
write when the ID is already on the private list (race protection).
- `push` filters snapshot entries by the private list and appends to a
per-workspace audit log `agnes-sessions-private-skipped.txt`.
- Queue format migrated from `<path>` to `<session_id>\t<path>`; legacy
one-column lines still parse (empty session_id, still upload, can't be
marked private retroactively — fine, they pre-date the feature).
- `install_claude_hooks` writes a workspace statusLine unless the user
already has a custom one (warn + preserve). Idempotent re-init.
- `install_claude_commands` ships `agnes-private.md` alongside
`update-agnes-plugins.md`. Per-template fallback so a missing template
doesn't get clobbered with the wrong content.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Fix setup-prompt + CLAUDE.md marketplace copy + drop skills step
Three issues against the post-PR-#240 / post-PR-#237 state:
1. Setup prompt's marketplace block trailer (both has-stack and
empty-stack variants) claimed the SessionStart hook keeps the
marketplace clone in sync via `agnes refresh-marketplace --quiet`
on every session and that admin grants land automatically — both
false since PR #237 (0.47.x) moved the install/update path out of
the hook into the `/update-agnes-plugins` slash command. The hook
is `--check`-only: detects server-side changes, prompts the user
to run the slash command, which does the full reconcile
interactively with output visible in the transcript.
2. The empty-stack variant framed composition as "admin grants only",
missing the actual three-source served stack:
(admin RBAC ∩ /marketplace subscriptions)
∪ system-mandatory plugins (admin-pinned, auto-applied)
∪ Flea market installs (skills/agents bundled, plugins standalone)
Updated copy spells out all three sources so analysts know where
their stack picks live, and what the SessionStart hook actually
does on change detection.
3. CLAUDE.md template's "Agnes Marketplace" section conflated
eligibility (`resolve_allowed_plugins` — what's listed) with served
stack (`resolve_user_marketplace` — what actually reaches Claude
Code). The two are different: a user can be RBAC-eligible for a
plugin without having subscribed to it on /marketplace. Rewrote
the section to distinguish the eligibility set from the served
stack and to describe the `--check`-only hook accurately.
Plus: deleted the setup prompt's interactive Skills step (final step
before Confirm). The named-opinion question — "do you want me to
bulk-copy every skill into ~/.claude/skills/agnes/ or pull on-demand
via `agnes skills show <name>`?" — had no obvious right answer for
new users at the tail end of a wall of technical steps. On-demand
lookup is the one-size-fits-all default; `agnes skills list/show`
remain discoverable and the CLAUDE.md template references specific
skills inline (e.g. agnes-data-querying in the BigQuery section)
where they're relevant. Layout: Confirm shifts from step 9 to step 8.
Tests updated, full setup/marketplace/welcome surface green (115
passed). Remaining full-suite failures are pre-existing (BQ/Keboola
fixtures, Windows charmap collection error in test_v26_keboola_e2e)
— verified against a clean stash, unrelated to this diff.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Fix session-queue race + snapshot PID-reuse data loss
Two blocker fixes from the PR #242 review:
1. Concurrent SessionStart hooks could corrupt the queue file on
Windows. Python's `open(path, "a")` is not atomic there — the CRT
does not pass FILE_APPEND_DATA to CreateFile, so concurrent
appenders (user opening several Claude Code windows simultaneously)
could interleave bytes mid-line. The malformed lines then silently
fail the parser and the entries are dropped.
Fix: wrap append_to_queue, requeue_failed, and snapshot_queue in a
short-lived FileLock on a dedicated `agnes-queue.lock`. Separate
from `agnes-push.lock` so capture-session hooks don't block on the
push command. New test_append_concurrent_threads_no_corruption
reproduces the race with 4 threads x 50 appends.
2. Snapshot filenames embedded only the PID (`agnes-sessions.snapshot.
<PID>.txt`). After a crashed push left a snapshot on disk and the
OS recycled the PID for a new push, `os.rename` would atomically
overwrite the recovery snapshot — every entry in it lost, silently.
Fix: append a uuid8 hex tail (`agnes-sessions.snapshot.<PID>.
<uuid8>.txt`). find_recovery_snapshots already globs the prefix
so it picks up both old and new format. New
test_snapshot_filename_is_unique_per_call asserts two consecutive
snapshots under the same PID don't collide.
Targeted tests green (47/47 in session_queue/capture_session/cli_push).
Full suite failures unchanged from baseline (pre-existing BQ/Keboola
fixture issues per CLAUDE.md).
* Auto-refresh workspace hooks + bash-wrap all hook entries (Windows)
Fixes from PR #242 second review (ZdenekSrotyr):
1. `uv.lock` regenerated to include `filelock 3.29.0` (declared in
pyproject.toml but missing from the lock file — CI's
lockfile-consistency check would fail; `uv pip install` on a clean
cache would silently miss the dep).
2. `agnes self-upgrade` now auto-refreshes the workspace Claude Code
hooks via the new `cli.lib.hooks.maybe_refresh_claude_hooks`. Closes
the silent-stop migration gap: a v0.48 workspace would auto-upgrade
the CLI from its existing SessionStart self-upgrade entry but never
pick up the new `agnes capture-session` SessionStart hook, leaving
the queue empty and `agnes push` uploading nothing.
The refresh fires on both the "info is None" fast path (CLI already
current — catches the second SessionStart after a prior upgrade)
and the install-success path. Guarded by `workspace_has_agnes_hooks`
so it never writes `.claude/settings.json` into directories that
aren't Agnes workspaces (e.g. `agnes self-upgrade` invoked from
`~/`). Errors are surfaced on stderr but never flip the upgrade exit
code.
3. All Agnes-managed hooks are now wrapped in `bash -c "..."`. The
self-upgrade+pull chained SessionStart entry was the only one still
shipping unwrapped — Claude Code on Windows runs hook commands
directly without a shell, so the `;` chain + `2>/dev/null` +
`|| true` shell syntax silently no-op'd on native Windows installs
without Git Bash on PATH. Workspaces still on the old form
auto-upgrade via the refresh path above.
Tests: +12 in test_lib_hooks.py (guard semantics, v0.48→v0.49
migration end-to-end, third-party-hook preservation, bash-wrap
invariant). +5 in test_self_upgrade.py (refresh fires on info=None,
fires on install success, skipped on failure, skipped on --check-only,
refresh failure never flips exit code).
130 targeted tests green. The 2 pre-existing Windows path-separator
failures in `test_smoke_test_detects_version_mismatch[uv|pip]` are
unrelated (path mismatch `\fake\uv\bin\agnes` vs `/fake/uv/bin/agnes`
in test asserts, pre-PR baseline).
* CHANGELOG: document PR-242 main features
Closes ZdenekSrotyr #4: the [Unreleased] block was missing entries for
the PR's primary surface — only the post-merge fix bullets and the
unrelated setup-prompt copy change were captured. Adds:
- ### Added: 6 bullets covering the session capture queue + new
`agnes capture-session` subcommand, `/agnes-private` slash + `agnes
mark-private`, `agnes statusline` + statusLine wiring, `--legacy-scan`
opt-in fallback, single-instance push lock, and the new `filelock`
runtime dep.
- ### Changed: BREAKING bullet on the SessionStart / SessionEnd hook
wire format change (capture-session as first SessionStart entry,
push self-heal removed, SessionEnd push detached via nohup, all
entries bash-wrapped). Folds the prior standalone bash-wrap bullet
into this consolidated entry — Z's review flagged the layout shift
as BREAKING, and grouping the related sub-changes makes the
migration story readable in one place.
- Operator migration is auto-handled by `maybe_refresh_claude_hooks`
invoked from `agnes self-upgrade` (separate Changed entry below).
No `agnes init` re-run required. Pre-queue session jsonls on
upgrading workspaces still need a one-off `agnes push --legacy-scan`
— flagged in the BREAKING bullet.
No code change; doc only.
* Drop permanent 4xx uploads instead of requeueing forever
Closes ZdenekSrotyr #5. Previously the push retry path requeued any
non-200 response except the literal "file not found on disk", so 401
(token expired), 403 (RBAC denial), 413 (payload too large), 400
(server-side validation) cycled through every push run forever — the
queue grew without bound and each run re-bombarded the server with the
same deterministically-failing upload.
Now 4xx (except 408 Request Timeout + 429 Too Many Requests, which the
HTTP spec marks as transient) is dropped and audit-logged to
`<workspace>/.claude/agnes-sessions-failed.txt`:
<iso_ts>\t<session_id>\t<status>\t<transcript_path>
5xx and network errors continue to requeue — those reflect server /
transport state that can change between runs, so retry is the right
behavior.
The audit log piggybacks on the push single-instance lock
(agnes-push.lock) — push is the only writer to this file, same as the
existing `mark_uploaded` and `mark_private_skipped` paths, so no
separate filelock is needed.
`agnes push --json` surfaces a new `dropped_permanent` counter; non-
quiet stdout mentions the audit-log path so operators tailing the
output have a pointer to the forensic trail.
Tests: +7 in test_cli_push.py (401/400/403/413 → drop; 408/429 →
requeue; 500/502/503 → requeue; network exception → requeue;
--json `dropped_permanent` counter; stdout audit-log pointer). +1 in
test_session_queue.py (mark_failed_permanent TSV format).
127/129 targeted tests green. The 2 pre-existing Windows
path-separator failures in `test_smoke_test_detects_version_mismatch
[uv|pip]` are unrelated (path mismatch `\fake\uv\bin\agnes` vs
`/fake/uv/bin/agnes` in test asserts, pre-PR baseline).
* Catch OSError in push lock acquisition
Closes ZdenekSrotyr #8. `acquire_or_skip` in `cli/lib/push_lock.py`
previously caught only `filelock.Timeout`. Any `OSError` from
`FileLock.acquire` — read-only filesystem, permission denied on
`.claude/`, disk full, hardware I/O failure — propagated as an
unhandled traceback.
Two visible failure modes:
- SessionEnd hook: `|| true` in the wrapper swallowed the error, so
daily pushes silently never ran. Operator had no signal.
- Manual `agnes push`: ugly Python traceback dumped to the terminal
instead of a clean exit.
Now `OSError` is treated the same as `Timeout` — yield `None`, caller
returns cleanly with rc=0. The operator's environment in these
scenarios has bigger problems than missing session uploads, so we
swallow rather than retry-loop or surface a noisy warning.
Test: `test_push_silent_exit_when_filelock_raises_oserror` patches
the `FileLock` used inside `push_lock` to raise OSError on acquire,
verifies push exits 0 with no traceback and the queue is preserved
for the next attempt.
* Address remaining S2 items from PR-242 review
Four items from ZdenekSrotyr's S2 list:
S2.10 — `_install_statusline` truthy check (cli/lib/hooks.py): replace
`if existing:` with explicit `if existing is None or existing == "":`.
Documents and tests the behavior for both edge cases (explicit-null
and empty-string `statusLine`) — both treated as "not configured"
rather than "explicit user opt-out", so we install ours. Two new
tests in test_lib_hooks.py pin the contract.
S2.6 — onboarding docs for /agnes-private. New "Private sessions"
subsection in `config/claude_md_template.txt` (next to Data Sync)
covering the slash command, statusbar indicator, and audit-log
location. One-line tip in `app/web/setup_instructions.py` so the
feature is discoverable at onboarding.
S2.9 — e2e privacy test (tests/test_e2e_privacy.py). Wires
capture_session → mark_private → push against a recording fake
api_post and asserts zero session uploads for the marked one.
Three cases: mark-before-capture (queue write skipped),
mark-after-capture (push-side filter catches it + audit-logs),
control (unmarked sessions upload normally).
David #8 — `--legacy-scan` help text now documents the
private-list gap (legacy entries carry empty session_id, so
the filter is not consulted). The practical impact is bounded —
pre-queue sessions cannot have been marked private since the
private list is a queue-era feature — but the disclaimer in the
help text means an operator running a backfill is not surprised.
68 targeted tests green (3 new e2e + 2 new truthy edge tests +
existing). 2 pre-existing Windows path-separator failures in
test_smoke_test_detects_version_mismatch[uv|pip] unchanged.
Remaining S2 items (statusline mkdir push-back, capture-session
silent-fail follow-up) handled in PR comment + follow-up issue
respectively.
* Address remaining S2 follow-ups (David #8, S2.7, David #11)
Three items left over from Mina's bbf63472 batch — that commit
addressed S2.6/S2.9/S2.10 + documented David #8 in help text but
deferred the actual implementations of S2.7, David #11, and the real
David #8 fix to follow-ups. This commit closes them.
David #8 — `agnes push --legacy-scan` now consults the private list.
Claude Code names jsonls `<session-id>.jsonl`, so the file stem IS
the session id; the legacy-scan path can apply the same private filter
the queue path uses. Both the dry-run and live-upload code paths fixed.
Help text updated (no longer warns the filter is bypassed). Two new
tests in test_cli_push.py cover the upload-skip path + the dry-run
`would_skip_private` segregation.
S2.7 — `statusline`/`is_private` no longer mkdir-pollutes arbitrary
workdirs. Split `_claude_dir` into `_claude_dir_writable` (used only
from `add_private`) and `_claude_dir_readonly` (no mkdir). The
read-only public helpers (`private_list_path`, `read_all_private`,
`is_private`) compose the no-mkdir variant by default; `add_private`
opts in via `writable=True`. Added a process-local mtime-keyed cache
around `read_all_private` so in-process callers (push doing one stat
per upload candidate, future `agnes diagnose`) don't re-parse the
file on every check. Cache eviction on `add_private` so a sub-second
write+read sequence doesn't see stale data even on coarse-mtime
filesystems. Two new tests pin the no-mkdir contract + the
in-same-second add+read consistency.
David #11 — `agnes capture-session` writes a breadcrumb log on every
invocation. New `<workspace>/.claude/agnes-capture-session.log` TSV:
`<iso_ts>\t<outcome>\t<detail>` where outcome covers every silent-
exit path (`ok`, `private_skip`, `empty_stdin`, `bad_json`,
`not_object`, `no_transcript_path`, `stdin_read_error`,
`write_error`). Gives operators a signal to detect "hook fires but
queue stays empty" — without it, an upstream Claude Code stdin-
contract change is invisible because the hook always exits 0. Log
rolls at 256 KiB so it doesn't grow unbounded on long-lived
workspaces. Best-effort: a breadcrumb-write failure is itself
swallowed so the hook contract stays "exit 0 always". Skipped in
non-Agnes workdirs (no `.claude/` exists) so opening Claude Code
in `~/` doesn't pollute it. Five new tests in test_capture_session.py
cover the success / bad_json / no_transcript_path / private_skip /
no-pollute paths.
115 targeted tests green (test_cli_push, test_capture_session,
test_private_list, test_session_queue, test_e2e_privacy,
test_lib_hooks, test_statusline, test_mark_private).
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-authored-by: ZdenekSrotyr <zdenek.srotyr@keboola.com>
* System plugin tier with mark/unmark fanout (schema v39)
Adds a mandatory plugin tier so admins can pin a small set of curated
plugins into every user's stack from day one. Marking a plugin via the
new toggle on /admin/marketplaces materializes resource_grants for every
group and user_plugin_optouts subscriptions for every user, so the
existing resolver pulls the plugin into every served set without a new
filter layer. Hooks on user-create (Google OAuth, magic-link, admin
POST, scheduler) and group-create propagate the same materialization to
new principals. UI locks: /admin/access disables the checkbox with a
SYSTEM pill; /marketplace cards swap the "In stack" green pill for an
amber "Required" badge with shield icon; the plugin detail install
button reads "Required by your org"; /my-ai-stack toggle is disabled.
Bypass paths return 409 (DELETE /api/admin/grants for system grants,
PUT /api/my-stack/curated/.../{enabled:false}, DELETE
/api/marketplace/curated/.../install). Unmark only flips the flag —
materialized rows persist so admins curate cleanup at their leisure
through the now-unlocked /admin/access checkboxes.
* Marketplace UX polish + drop legacy /store and /my-ai-stack pages
Two-part cleanup post-v39:
(1) Page deletion. /store and /my-ai-stack were already replaced by
/marketplace?tab=flea and /marketplace?tab=my respectively, but the
standalone routes lingered. Hard delete in dev mode — no redirects,
stale bookmarks 404. The /store/new upload wizard, the flea
detail/edit pages, the admin queue, and all /api/store/* +
/api/my-stack endpoints (CLI consumers) stay. Internal hardcoded
hrefs in the upload wizard's Cancel button and the advanced-setup
page repointed to the marketplace tabs.
(2) Detail-page install button rework. The single button that morphed
between "+ Add to my stack" and "✓ In your stack" did not
communicate uninstall affordance. The installed state now renders an
inline white status label *before* a separate red-bordered
"✕ Remove from stack" button on the same row, both at identical
height to avoid layout shift. System plugins keep their locked amber
"✓ Required by your org" pill (no Remove button — API refuses 409).
The post-action hint panel now fires on remove too with the title
flipped to "✓ Removed from your stack" — Claude Code needs the same
/update-agnes-plugins refresh either way.
Also: /admin/marketplaces Details modal "Mark as system" toggle
redesigned. The button was near-invisible (matched neutral row
metadata). It's now a balanced amber-toned chip with shield icon
and a structured confirm modal replacing the native confirm() dialog
that summarizes fanout consequences before commit.
* Move stack-hint inside hero with glass-on-gradient styling
The post-action hint card ("✓ Added to your stack" with the
/update-agnes-plugins recipe) used to live below the hero in
panel-what (gray card on white page body). Clicking add/remove
inserted/removed it between the hero and content, shifting the
panels below — a noticeable scroll jump.
The hint is now anchored inside the hero's top-right corner alongside
the install/remove buttons, both as flex children of an absolutely
positioned .actions container. The card uses a translucent
white-on-glass treatment that adopts the hero's kind color (blue for
plugin, green for skill, purple for agent) without per-kind branching.
Hero is always tall enough (160px photo) to contain the action+hint
stack without overflow, so toggling the hint visibility doesn't grow
the hero or shift body content.
The hero-head grid reserves a third 300px column for the absolute
actions overlay so meta gets the proper 1fr free space instead of
being squeezed by a padding-right hack. Responsive breakpoint at
1100px reflows the actions stack below hero-head when the viewport
isn't wide enough to keep meta + actions side-by-side comfortably.
* Add optional -DataPath bind mount to run-local-dev.ps1
When the operator wants to inspect DuckDB files (system.duckdb, extracts,
marketplaces, store/, …) directly from Windows Explorer, the named volume
inside the Docker Desktop WSL VM isn't reachable. The new -DataPath param
generates a transient compose override that rebinds /data on app, scheduler,
extract (and Caddy's /srv:ro mirror) to a Windows host folder.
Fully additive — when -DataPath is omitted everything behaves exactly as
before: no override file is generated, $composeFiles array is unchanged,
finally cleanup is a no-op. Existing positional invocations
(.\run-local-dev.ps1 up | down | logs) keep binding to $Action because
$DataPath is a named-only parameter with no Position attribute.
The override is written via [System.IO.File]::WriteAllText so the YAML is
BOM-less across PS 5.1 / 7+ — Compose rejects BOM-prefixed YAML on Windows.
The override file is unique per PID and removed in the script's finally
block so concurrent invocations and crashes don't leak files.
* factor mark_system fanout into UserCuratedSubscriptionsRepository
The endpoint imported UserCuratedSubscriptionsRepository, ignored it
(noqa: F841), then duplicated the user-side fanout SQL inline. Adds
fanout_system_for_plugin() symmetric to the existing
fanout_system_for_user() and routes mark_plugin_system through it —
removes the dead import + 14 lines of inline SQL, returns the same
`affected_users` delta count, no behavior change.
* drop customer-specific path from .ps1 example
Per CLAUDE.md vendor-agnostic OSS rule: replaced
C:\\Business\\Groupon\\Agnes\\agnes-data with the generic
C:\\Users\\<you>\\agnes-data placeholder so the docstring
example reads cleanly on any reviewer's box.
* release: 0.48.0 + parallelize Release-workflow pytest
Cuts the release shipped via #228#230#231#232#233#234#236#237#238#239#240 plus this PR (#241). Major changes:
- System plugin tier (schema v39) — admins mark a plugin mandatory; fans
out RBAC grants + subscriptions to every existing user/group plus
hooks for new principals
- BREAKING: removed standalone /store + /my-ai-stack page routes
(replaced by /marketplace?tab=flea + /marketplace?tab=my)
- Setup-prompt + bootstrap recovery fixes (#240)
- DuckDB CHECKPOINT-on-shutdown + 60s compose grace (#235)
- Marketplace + flea-market UX polish, agnes-metadata.json enrichment
Bonus: switch release.yml test step to `-n auto` (matches ci.yml).
Single-threaded was 15-20 min and frequently the bottleneck on PR
mergeability — now ~6 min.
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
Co-authored-by: ZdenekSrotyr <zdenek.srotyr@keboola.com>
* Setup-prompt + bootstrap fixes from David's 2026-05-10 init report
Three issues from clean-machine bootstrap evidence:
1. `agnes refresh-marketplace --bootstrap` failed to recover when the
local clone existed but Claude Code's marketplace registry had lost
the `agnes` entry. Bootstrap path now parses
`claude plugin marketplace list`, re-runs
`claude plugin marketplace add ~/.agnes/marketplace` when missing,
and treats `add` failures as fatal (was warn-and-continue, root cause
of the cascade into "Marketplace 'agnes' not found" plugin install
errors).
2. Setup prompt now always emits the marketplace-registration block,
even when the operator has zero plugin grants. Pre-wires the
SessionStart hook so future admin grants land automatically without
re-running setup. Block copy adapts: empty list shows
"no plugins granted yet", populated list shows "install plugins".
3. Setup prompt registers the Atlassian Remote MCP server unattended
(`claude mcp add --transport sse atlassian
https://mcp.atlassian.com/v1/sse`). Hosted Remote MCP, OAuth handled
automatically by Claude Code on first use. Asana / GWS stay on the
/home connector cards (PAT/keychain flows don't fit unattended
bootstrap).
Confirm step nudges the user toward the /home connector cards for the
PAT-flow services. CLAUDE.md template renames the marketplace section
to "Agnes Marketplace" and documents that all plugins are addressed as
`<plugin>@agnes` regardless of upstream slug.
Layout: Confirm shifts from step 6/8 to step 9 across all variants
(preflight, marketplace, MCP all unconditional). Tests updated.
* Link Claude license options from /home install pane
Step-1 Claude install on /home pointed users to OAuth without
explaining what to do if they don't have a Pro/Max subscription. Add
a one-line follow-up link to the plan-tier section on /setup-advanced
(new `#claude-plan` anchor) so first-time users discover the
subscription tiers rather than bouncing on the OAuth screen.
* Add idempotent + no-TLS-bypass guardrails to /home connector prompts
The Asana / Google Workspace / Atlassian connector prompts on /home
already shipped a precheck step that short-circuits when the service
is already wired, but they didn't carry the same idempotency +
surface-errors-verbatim + don't-disable-TLS-verification guardrails
the bash bootstrap prompt has. Add a one-paragraph 'Ground rules'
block at the top of each prompt so a connector failure doesn't
tempt the model into bypass workarounds, matching the same posture
David's 2026-05-10 init report flagged for the bash flow.
* skip Source: lines in marketplace registry detector
`claude plugin marketplace list` prints a `Source: <local path>` line
under each registered marketplace; the local clone almost always lives
under a path containing the marketplace name itself
(`~/.agnes/marketplace`). A naive \\bagnes\\b match over the full
stdout therefore false-positives whenever ANY unrelated marketplace
sits under `~/.agnes-…/` or similar. Filter Source: lines out before
matching so the recovery path actually re-adds when needed instead of
silently falling through to a broken `marketplace update agnes`.
Adds regression test covering the substring-only case.
* drop customer-specific tokens from CHANGELOG entries
Per CLAUDE.md vendor-agnostic OSS rule ("nothing customer-specific
... in changelogs"):
- "agnes-vrysanek.groupondev.com" -> "a private-CA Agnes deployment"
- "Groupon Marketplace / groupon-marketplace" -> "<Org> Marketplace /
<org>-marketplace" (placeholder example)
- Removed "David flagged" attribution language; init-report context
stays intact, just stripped of the named host + brand
---------
Co-authored-by: ZdenekSrotyr <zdenek.srotyr@keboola.com>
* feat(store): flea-market entity edit feature with version history (schema v38)
Owner + admin can now edit a store entity from a real Edit page at
/marketplace/flea/{id}/edit, replacing the prior "coming soon"
placeholder. Editable: display name, description, category, video
URL, cover photo, and an optional new bundle. Type is locked (400
type_locked). Display-name change renames the on-disk slug for both
live plugin/ and version dirs (reuses rename-on-archive helper).
Schema v38 (originally drafted as v37; renumbered after rebase onto
main where v37 was taken by the curated marketplace enrichment).
Versioning model:
* Each bundle update bakes into ${DATA_DIR}/store/<id>/versions/v<N+1>/plugin/
and runs the standard guardrails pipeline.
* DEFERRED PROMOTION: live plugin/ + entity.version_no stay at the
prior approved version through the LLM review window so existing
installers keep receiving the previously approved bundle. Live swap
+ version_no/version/file_size bump happen only on LLM approval.
Blocked verdicts leave the prior version serving forever.
* store_entities gains version_no INTEGER + version_history JSON.
Each version_history entry carries hash, sha256, size, submission_id,
created_at, created_by.
* Existing entities backfill to v1 with a single-entry history seeded
from the row's current `version` hash. Initial create also seeds
versions/v1/plugin/ so future restore can copy v1 bytes forward.
Concurrency:
* Block-while-pending: an in-flight LLM review blocks any further edit
with 409 prior_version_pending. Owner waits 5-30s; Edit button on
detail page renders disabled in the same window via the new
edit_in_flight flag (decoupled from quarantine_sub since the
deferred-promotion flow keeps visibility='approved').
Rollback:
* New endpoint POST /api/store/entities/{id}/versions/{n}/restore
(owner + admin). Copies vN bundle forward as v<max+1> and re-runs
guardrails (rules tighten over time; pre-approved bundles re-validate).
Forward-only history. Same deferred-promotion semantics — live stays
at prior version until LLM approves the restored copy.
UI:
* New /marketplace/flea/{id}/edit page (owner + admin gated).
* Versions card on plugin + item detail templates (owner/admin only)
via shared _flea_versions.html partial.
* Admin queue gains v# column with current badge + separate Hash
column. Submission detail surfaces Version + Bundle hash rows.
* Activity timeline split into per-submission + entity-wide cards;
entity-wide rows render vN chips when audit row params reference
a specific version.
* Section headers (Manifest / Static / Quality / LLM review) tag
with vN chip via shared macro.
* Reviewed-by-model field surfaces explanatory text per status.
* Banner upload-failure now redirects to detail page on
submission_blocked instead of staying stuck.
Tests: 24 in tests/test_store_entity_versions.py covering metadata-
only edit, bundle-edit version bump, type lock, block-while-pending,
name change disk rename, restore flow + 404/400/403 paths, edit page
404 for non-owner, versions card visibility gating, admin queue v#
column, admin detail Version/Hash rows, deferred-promotion installer
contract (pending review doesn't break installer / blocked verdict
keeps prior / approved promotes), admin can edit/restore non-owned,
restore deferred promotion, audit log per-version params. 214 tests
green across guardrails + edit + admin + repo + schema suites.
* docs(store): refresh update_entity docstring to match deferred-promotion + submission-status gate
Bring the docstring in sync with the actual fixes from the prior
commit. The pre-fix wording said the gate read
visibility_status='pending' AND submission status — under deferred
promotion that would never fire for v2+ edits. Now describes:
- Block-while-pending gates on submission.status DIRECTLY,
independent of visibility (so v2+ deferred-promotion edits don't
slip through).
- Display-name + bundle change defers the live rename to promotion;
metadata-only renames stay immediate.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Move marketplace plugin updates from hook to /update-agnes-plugins skill
The SessionStart hook used to run `agnes refresh-marketplace --quiet`,
which performed a full fetch+reset+install cycle on every Claude Code
session start. That work was invisible to the user, slowed session
startup, and was unrecoverable interactively when something failed.
Split the responsibility:
- `agnes refresh-marketplace --check` is a new lightweight detector:
`git fetch` only, compares local HEAD with remote FETCH_HEAD, emits
a Claude Code hook JSON message pointing the user at
`/update-agnes-plugins` when the marketplace has changes. No reset,
no plugin install/update side effects.
- `/update-agnes-plugins` is a new slash command (installed by
`agnes init` into `<workspace>/.claude/commands/`) that runs
`agnes refresh-marketplace` (default chatty path). Output streams
into the Claude Code transcript so the user sees install/update
progress and can react to errors interactively.
- The SessionStart hook now runs `--check`. Existing workspaces
auto-upgrade on next `agnes init` (substring marker matches both
the old `--quiet` entry and the new `--check` one).
BREAKING: `agnes refresh-marketplace --quiet` is removed. Old hooks
calling it silent-noop after the CLI upgrade (the hook's `|| true`
swallows the unknown-flag error) until re-init rewrites them.
* Point marketplace 'Added to your stack' hint at /update-agnes-plugins
The post-install green panel on plugin and skill/agent detail pages
referenced the SessionStart auto-install path and a shell-prompt
`agnes refresh-marketplace` invocation. With the hook now being
detect-only, that copy was misleading — the actual install path is
the new slash command.
Condensed to a single instruction: "Open a new Claude Code session
and run:" followed by `/update-agnes-plugins` in a copy-chip.
JS clipboard string updated to match.
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
The marketplace step (step 5) emitted `git config --global
http.<host>/.sslVerify false` on AGNES_DEBUG_AUTH=1 instances when no
ca_pem was readable from AGNES_TLS_FULLCHAIN_PATH. Two problems:
1. Claude Code auto-mode classifiers ("do not disable TLS verification"
rule) block the line, breaking hands-free setup.
2. It silently masked operator misconfiguration — a debug-auth instance
without a fullchain on disk fell through to a TLS-disabled clone
instead of surfacing the missing cert.
After the cross-platform trust block (#137), self-signed and private-CA
servers are fully covered by step 0 reading the fullchain via
_read_agnes_ca_pem; publicly-trusted certs need no bootstrap at all.
The legacy fallback no longer covers a real scenario — verified by
running step 5 without sslVerify=false against a self-signed instance.
BREAKING: drops `self_signed_tls` parameter from
app.web.setup_instructions.resolve_lines and render_setup_instructions
(only consumed by the deleted block). The AGNES_DEBUG_AUTH env var
itself is unchanged — still gates /api/me_debug and the dropdown link.
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
* Curated marketplace enrichment via agnes-metadata.json + curator metadata
Adds a second well-defined metadata file `.claude-plugin/agnes-metadata.json`
that upstream marketplace repos can opt into, providing per-plugin (and
per-skill / per-agent) cover photo, demo video URL, doc links, and
category override. The Claude Code marketplace contract is untouched —
agnes-metadata.json + the convention `.agnes/` directory are stripped
from the synthetic Claude Code marketplace served via /marketplace.zip
and /marketplace.git/*, so user instances see a clean Claude Code repo
with no Agnes-only metadata.
Highlights:
- DB schema v32 — adds curator_name + curator_email on marketplace_registry,
cover_photo_url + video_url + doc_links on marketplace_plugins.
- Mandatory curator at marketplace registration, editable later through
the admin UI; surfaces on cards + detail pages in place of owner_todo.
- External-asset mirror cache at ${DATA_DIR}/marketplace-cache/<slug>/
with conditional GET, 60s timeout, 10 MB body cap, SSRF guards, and
Wikipedia-policy-compliant User-Agent.
- Strict drop semantics — anything Agnes can't deliver as a real PDF /
Markdown / plain text doc, or a real PNG / JPEG / WebP cover, is
dropped from the served metadata; UI looks identical to no-entry case
(gradient placeholder for missing covers, no row in the doc list).
- Doc allowlist + image allowlist enforced on both the curated mirror
flow and the Flea upload flow (/store/new); shared module
src/marketplace_assets.py.
- New /api/marketplace/curated/{mp}/{plugin}/{asset,doc,mirrored}/...
endpoints with path-traversal guards + RBAC + Content-Disposition
attachment for docs.
- Curator-focused format guide at /marketplace/format-guide; canonical
source is docs/curated-marketplace-format.md, also linked from the
admin /admin/marketplaces page next to + Add Marketplace.
See CHANGELOG.md under [Unreleased] for the full breakdown.
* Fix format-guide test assertion to match shortened disclaimer
The 'Flea Market' phrase was trimmed out of the disclaimer in
docs/curated-marketplace-format.md after the curator-focused rewrite.
Update the rendered-HTML test to assert the channel-scoping phrase
that's actually present ('Curated Marketplace channel only') rather
than the 'Flea Market' contrast that's no longer in the doc.
* Drop unused 'version' field from agnes-metadata.json schema
The parser never read it; it was a YAGNI placeholder for future
schema evolution. Curators don't need to wonder what to put there
when adding the file for the first time. Will be re-added if and
when we actually introduce a backwards-incompatible schema change.
* Harden asset mirror against SSRF via redirect + DNS rebinding
The pre-flight _is_safe_url check validated only the initial URL;
urllib.request.urlopen then followed redirects and re-resolved DNS for
the actual connection — both bypassable. Attacker-controlled origin
could 302 to http://169.254.169.254/... and exfil cloud metadata;
attacker-controlled DNS could return public IP first / 127.0.0.1 second.
Replace urlopen call with a shared OpenerDirector wired through three
custom handlers: _SafeRedirectHandler re-runs SSRF allowlist on every
redirect Location (max 5 hops, down from urllib's 10), and
_PinnedHTTPHandler / _PinnedHTTPSHandler connect to the IP that passed
validation rather than re-resolving the hostname. TLS SNI + cert verify
stay bound to the original hostname.
_resolve_safe returns the validated IP (the existing _is_safe_url
2-tuple wrapper stays for backwards compatibility) and rejects round-
robin DNS that mixes a public + private record. _UnsafeRedirectError
is a typed exception so _fetch_url can map redirect blocks to terminal
'rejected' status (not transient 'failed'). _http_open is the single
call site so tests can mock at one well-defined seam.
Tests cover redirect blocking (link-local, loopback), redirect-error
unwrapping inside URLError, pinned-IP connection target, and the
end-to-end DNS-rebinding scenario. Existing tests that mocked
urllib.request.urlopen are migrated to mock _http_open.
* Harden /asset/ endpoint against stored XSS
The endpoint served any file in the cloned marketplace repo with
stdlib-detected Content-Type, so a curator who landed evil.html (or a
renamed evil.png carrying HTML bytes) in the working tree got a
same-origin XSS — the response shares cookie scope with /admin and
/api/me/*.
The asset endpoint is image-only by contract (cover photos referenced
from agnes-metadata.json + inner skill / agent cards), so applying the
same allowlist + magic-bytes pattern that /doc/ already uses closes
the gap without breaking any legitimate use case. Three layered
checks: extension in IMAGE_EXTENSIONS (.png/.jpg/.jpeg/.webp; SVG
excluded — <script> inside SVG executes), validate_image_file magic
bytes (defeats rename-extension attack), Content-Type pinned from the
validated extension (never stdlib mimetypes).
Defense-in-depth: X-Content-Type-Options: nosniff stops browser MIME
sniffing; Content-Security-Policy: default-src 'none' blocks script /
iframe execution even if a future regression let HTML through.
Tests cover the .html extension reject, the renamed-HTML-as-PNG magic-
bytes reject, the .svg reject, and the happy-path PNG with security
headers attached. The pre-existing path-traversal test seeds a real
PNG instead of ok.txt now that the endpoint is image-only.
* Enforce mandatory curator on marketplace PATCH
The POST handler enforced curator_name + curator_email at create time,
but PATCH treated empty / missing curator inputs as 'no change'. Legacy
rows that pre-date v32 (curator_name=NULL) could be edited indefinitely
without ever filling the curator gap, and OWNER_TODO_PLACEHOLDER lingered
on every /marketplace card.
Reject the PATCH with 400 when the post-merge row would persist with
empty curator. The check fires after the existing field-merge logic, so
once-filled rows that don't touch curator still pass through (their
existing values fall through from the DB row). DB column stays nullable
so untouched legacy rows continue to coexist — the gate fires only the
moment an admin opens the edit modal.
Existing PATCH semantics preserved: empty-string input still means 'leave
existing value alone', and once-filled curator can't be cleared (those
test cases pass unchanged). New test seeds a legacy row directly via the
repository, then exercises url-only PATCH (rejected), partial-fill PATCH
(rejected), and full-fill PATCH (succeeds); a follow-up no-curator PATCH
on the now-formed row also passes.
* Drop unused curated-marketplace helpers (PR #234 review)
* build_db_payload — imported by src/marketplace.py but never called.
The strict-drop semantics it would have implemented were re-written
inline in _refresh_plugin_cache (see the comment block there). The
standalone helper still carried the old fall-back-to-original-external-
URL-on-mirror-failure behaviour, which contradicts the documented
drop-when-can't-deliver contract — a future contributor who re-wired
it would have introduced a silent regression. Delete with the helper
+ the import + the comment that referenced it.
* _resolve_marketplace_name — one-line shim with no remaining call
sites. Callers use _resolve_marketplace_meta which returns name +
curator together, avoiding the double DB hit the shim exists to
hide.
* '# noqa: F401 Optional kept for forward-compat' was wrong — Optional
IS used in src/marketplace.py (line 70 and line 238). Drop the noqa
comment so a future ruff run doesn't try to remove a real import.
Removing build_db_payload also drops the only remaining use of Optional
in src/marketplace_metadata.py, so the import comes out there too.
* Cap agnes-metadata.json size + catch RecursionError on parse
The reader is invoked once per marketplace per sync and the file is
curator-controlled. Two failure modes were unguarded:
* Multi-GB JSON: path.read_text() pulled the whole file into memory
before json.loads even ran. A curator with commit access to an
upstream repo could OOM the sync worker.
* Deeply-nested JSON under any size cap: cpython's recursive object /
array parser raises RecursionError at ~1000 levels of depth.
RecursionError is a RuntimeError, not ValueError, so the existing
catch let it propagate up and abort the entire sync — every other
marketplace in the same pass got skipped.
Add AGNES_METADATA_MAX_BYTES = 1 MiB (a real metadata file with covers,
docs, categories for ~50 plugins fits in <100 KB so the cap is
generous) and gate the size check on path.stat().st_size before the
body read. Broaden the parse except to (ValueError, RecursionError)
with a unified log line. Both failure modes degrade to the same
empty-dict fall-back the malformed-JSON path already used, so one bad
upstream never aborts the rest of the sync.
Tests cover the size cap firing before json.loads (whitespace-padded
valid JSON exceeding the cap) and the recursion path (5000 nested
arrays — past cpython's default recursion limit but well under the
size cap).
* Persist asset-mirror manifest per body write, before unlink
sync_assets wrote each body atomically (tmp + rename) but persisted
the manifest only at the end of the batch. A kill -9 mid-Phase 2 left
on-disk files the manifest never referenced. Once a curator dropped
that URL from agnes-metadata.json, Phase 3's cleanup had no record of
the file and the orphan stayed forever — there's no GC pass walking
the cache dir today, so disk would slowly bloat.
Phase 2 (body-write iteration): after the in-memory manifest mutation,
persist BEFORE unlinking the previous body. The crash window narrows
from 'all of Phase 2' to 'between persist and unlink' (microseconds).
A persist failure mid-batch keeps the previous body on disk — the on-
disk manifest still references it, and a stale-but-existing file beats
a 404. Cost: one extra tmp+rename per body write; manifest is a few KB
so the overhead is negligible vs. the HTTP fetches.
Phase 3 (curator-removed URLs): same discipline. Collect the to-delete
relpaths, persist the manifest with the entries already gone, THEN
unlink. A crash mid-cleanup leaves at most a microsecond window where
files exist despite the manifest no longer naming them. The next sync
reads the (correct) manifest and the orphan stays orphaned, but the
served state is consistent.
Tests cover per-body persist call count, the post-update on-disk
manifest content, and Phase 3 ordering verified by reading the on-disk
manifest from inside Path.unlink.
* Consolidate marketplace video embeds + format-guide CSS
The YouTube nocookie / Vimeo / <video> / link-fallback detection logic
was duplicated verbatim in marketplace_plugin_detail.html and
marketplace_item_detail.html (~40 JS lines each, with subtly-different
inline styles). Both templates now {% include %} a single
_marketplace_video_embed.html partial inside their IIFE so the regex,
the nocookie attribute set, and the unknown-host link fallback live in
ONE place — future tweaks (new host, new attribute, fixed sandbox flag)
no longer need to be applied twice in lockstep.
The .video-wrap selectors (one inline <style> rule in plugin_detail,
one inline style='...' attribute in item_detail) are replaced by the
existing .video-embed 16:9 wrapper in style-custom.css, with new
.video-embed video / .video-embed a child rules added so the wrapper
handles all four embed shapes uniformly without per-template
positioning.
The 60-line inline <style> block in marketplace_format_guide.html
moves verbatim to style-custom.css under a new 'Marketplace format
guide page' section, scoped to .format-guide so other pages aren't
affected.
No user-visible behaviour change: the rendered HTML for valid
YouTube / Vimeo / mp4 / external links is byte-identical to before,
and the format-guide page renders the same.
* Maintainability cleanup batch (PR #234 review)
#10: drop _path_under from app/api/marketplace.py — it was a byte-
equivalent clone of _safe_join (same Path.resolve(strict=True) +
relative_to() containment check). The three v32 endpoint handlers
(/asset, /doc, /mirrored) now share the existing helper.
#14: rename src/marketplace_assets.py → src/marketplace_asset_validation.py
so the file's purpose is obvious from the name and the previous
overlap with src/marketplace_asset_mirror.py is gone. Six call-site
imports updated in lockstep; CHANGELOG references under [Unreleased]
updated to track the new path.
#11: consolidate the URL builders that resolve
/api/marketplace/curated/<slug>/<plugin>/{asset,doc,mirrored}/...
paths. _internal_asset_url / _internal_doc_url / _mirrored_asset_url
lived in src/marketplace.py, while a copy named _mirrored_url lived
in app/api/marketplace.py with a 'must stay aligned' comment. New
module src/marketplace_urls.py is the single source of truth — both
call sites import from it and a future URL-format tweak only needs
to change one file. The _ROUTE_PREFIX constant collapses the per-
function f-string repetition. The route-handler endpoints themselves
still own the path string literals (keeping the builders identical
to the route declarations remains a checklist item, not a runtime
guarantee).
* Re-key asset-mirror manifest by (plugin, url) + dedup HTTP fetches
The manifest used to be keyed by URL alone, so two plugins in the
same marketplace referencing the same external image (a shared CDN
icon, a common cover) collided on entry.plugin_name — last writer
won. The DB row for the losing plugin then stored a served URL
pointing under the winning plugin's tree, and require_resource_access
denied legitimate access on one side and let the other plugin's user
reach the wrong asset.
In-memory: Dict[Tuple[str, str], MirrorEntry] keyed (plugin_name, url).
On disk: format flips from {url: entry} dict to [entry, ...] list of
self-describing entries (each carries plugin_name + url + the
previous fields). JSON keys can't be tuples; encoding 'plugin::url'
would just shift the parsing burden.
Phase 1 of sync_assets deduplicates fetches by URL — three plugins
sharing one URL share one HTTP request. The conditional-GET prior is
picked from any owning plugin's prior entry; if their etags diverge
(rare) we miss one 304 and pay for a full re-download instead.
Phase 2 still creates a per-(plugin, url) manifest entry pointing
under the plugin's own subdir, and Phase 3 cleanup is keyed the same
way so dropping a URL from one plugin's metadata doesn't disturb
another plugin still referencing it.
Body files stay per plugin (RBAC-clean isolation: deleting plugin A's
cache can't strand plugin B). Bandwidth saved by fetch dedup.
Consumer code re-keyed: src.marketplace._refresh_plugin_cache rebuilt
served_url_for / mirror_status as composite-keyed maps;
app.api.marketplace._resolve_external_via_mirror /
_curated_inner_cover / _curated_inner_enrichment look up by
(plugin_name, url).
Tests cover per-plugin manifest entries with shared URL, the single
HTTP fetch for N plugins, and Phase 3 drop-one-keep-other. All
existing tests migrated to composite key access; v2 list format
assertions verify on-disk shape.
* Migrate asset mirror from urllib.request to httpx
The asset mirror was the only HTTP call site in Agnes still using
urllib.request; every other module (CLI, Jira / OpenMetadata / OpenAI
connectors, scheduler, Telegram bot) already used httpx. The asset
mirror was added in this PR's base commit, so this is the only chance
to bring it into convention before someone copies it as 'the pattern
for HTTP fetches in Agnes'.
Three concrete benefits beyond consistency:
* SSRF defence collapses from five urllib classes
(_PinnedHTTPConnection, _PinnedHTTPSConnection, _PinnedHTTPHandler,
_PinnedHTTPSHandler, _SafeRedirectHandler) into one
_SSRFGuardTransport. httpx invokes handle_request() on every redirect
hop, so re-validation is free — we don't need a custom redirect
handler at all.
* DNS-rebinding defence: the transport rewrites request.url.host to the
SSRF-validated IP before delegating to super().handle_request().
httpcore connects to whatever URL.host says, so this pins the
connection without subclassing HTTPSConnection. The original hostname
goes into the Host header + the sni_hostname extension so TLS / vhost
routing still bind to the curator-supplied hostname.
* Error handling: one httpx.HTTPError catch-all for transport errors,
plus specific httpx.TimeoutException / httpx.TooManyRedirects branches
for clearer diagnostics. Matches the _translate_transport_error shape
in cli/client.py.
The shared httpx.Client is built lazily at module load (same pattern as
cli/client.py:_get_shared_client) with follow_redirects=True,
max_redirects=5, timeout=HTTP_TIMEOUT_SEC, and our custom transport.
Externally observable behaviour is unchanged: same FetchOutcome
statuses, same manifest format, same conditional GET semantics, same
body-size cap.
Tests migrated from urllib-shaped fakes to httpx-shaped (status_code,
iter_bytes, context manager). Five urllib-specific tests replaced with
httpx equivalents — three transport unit tests + one DNS-rebinding
integration test that verifies host rewrite via monkey-patched
super().handle_request. One test deleted without replacement
(unwrap-URLError-wrapping-an-_UnsafeRedirectError — urllib-specific,
not applicable to httpx).
* Surface curated agnes-metadata enrichment on My Stack tab
GET /api/marketplace/items?tab=my built each curated row from the
on-disk marketplace.json by way of resolve_allowed_plugins, which
doesn't carry the agnes-metadata enrichment columns
(cover_photo_url, video_url, category override, doc_links). The
handler then hard-coded cover_photo_url=None on the synthetic row.
Result: once a user clicked '+ Add to my stack' on a curated card,
the same plugin in tab=my rendered with the gradient placeholder
instead of its cover photo — confusing parity break vs. the curated
tab where the same row goes through MarketplacePluginsRepository
and gets the enriched columns.
Pre-load the enriched marketplace_plugins rows for every marketplace
the user is subscribed to, then look each granted+subscribed plugin
up by (marketplace_id, plugin_name). Fall back to the on-disk
synthetic shape only when the DB row is missing — happens during
the rare race where RBAC is granted before the first sync cycle
ingests the plugin. RBAC gating (granted set from
resolve_allowed_plugins) is unchanged so this fix can't widen
visibility; it just upgrades the data shape behind cards the user
was already going to see.
Per-marketplace list_for_marketplace beats N gets — typical user is
subscribed to <5 marketplaces, so this is at most a handful of
queries vs. one per subscribed plugin.
Regression test seeds a plugin with cover_photo_url + category
override, subscribes the user, hits /api/marketplace/items?tab=my,
and asserts photo_url + category come through. The misleading
'fall through to gradient until the user re-visits the curated tab'
comment is gone.
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
The original list-form _V34_TO_V35_MIGRATIONS ran four ALTER
statements in sequence:
ADD _vis_v35 → UPDATE _vis_v35 = visibility_status →
DROP visibility_status → RENAME _vis_v35 TO visibility_status
If the RENAME failed for any reason after the DROP succeeded — DuckDB
lock contention at startup, scheduler-vs-app race opening
system.duckdb, container kill mid-migration, etc. — the DB was
stranded with _vis_v35 populated and visibility_status missing. The
schema_version row never bumped because the UPDATE at the bottom of
the migration ladder runs only when every step succeeded. Subsequent
restarts then hit DROP visibility_status again with no IF EXISTS
guard and looped on the same error; the only recovery was hand-
editing the DB.
Replace the list with a Python function _v34_to_v35_migrate that
inspects the table's columns up front and dispatches into one of
three paths:
* clean v34 (visibility_status present, _vis_v35 absent) — run the
full rebuild
* partial v35 (_vis_v35 present, visibility_status absent) — finish
the RENAME alone, data is already in _vis_v35 from the prior
UPDATE
* both columns present (rare; aborted before DROP) — drop the temp
and keep visibility_status
The audit columns (archived_at, archived_by) ship first behind
IF NOT EXISTS so they're safe in all states. Operators stranded by
the original bug now recover automatically on next startup.
Tests cover the three direct paths plus an end-to-end scenario where
_ensure_schema walks a schema_version=32 DB with the half-applied
state up through to v36.
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
* feat(store): flea-market upload guardrails + soft delete + JOIN-based admin queue
Adds an end-to-end guardrails pipeline for store uploads (manifest +
static-security + LLM review), persists blocked bundles for forensics,
introduces soft-delete (Archive) semantics, consolidates the legacy
/store/{id} surface into /marketplace/flea/{id}, and reworks the admin
queue so lifecycle filters read live entity visibility via LEFT JOIN
rather than a denormalized submission column.
Schema v29 → v35:
* v29 store_submissions table + store_entities.visibility_status
* v30 file_size, bundle_sha256, bundle_purged_at on submissions
* v31 reshape store_submissions (drop legacy unique on entity_id)
* v32 store_entities.archived_at/by + 'archived' visibility value
* v33 drop store_submissions.retry_count (unused)
* v34 ensure idx_store_submissions_entity exists post column-drop
* v35 broaden visibility_status enum + JOIN architecture cutover
Pipeline (src/store_guardrails/):
* Inline checks: manifest_check, static_scan, quality_check
* LLM review configurable haiku|sonnet|opus (default haiku)
* BackgroundTasks-driven async path with structured-output JSON
* Per-submitter daily quota (default 50)
* 30-day TTL purge job (POST /api/admin/run-blocked-purge)
* Bundle SHA256 + size persisted; sha256 survives purge for forensics
Visibility model:
* pending | approved | hidden | archived
* _enforce_visibility returns 404 (no leak) for non-owner non-admin
* Owner sees own non-approved entries via include_owner_id widening
* Install refused with 409 entity_not_approved when not approved
Soft-delete (DELETE /api/store/entities/{id}):
* Default = soft (visibility_status='archived'); existing installs
keep getting served the bundle so users don't lose the plugin
* ?hard=true admin-only: drops bundle + cascades user_store_installs
* Hard-delete preserves entity_id on submission as tombstone so
audit_log linkage survives for the activity timeline
Admin queue lifecycle (the JOIN refactor):
* Verdict (store_submissions.status) is immutable forensic record
* Lifecycle (store_entities.visibility_status) is live state
* /admin/store/submissions Archived chip translates to
`e.visibility_status='archived'` via LEFT JOIN — any path that
flips visibility surfaces in the queue immediately
* Detail page renders Status (verdict) and Entity lifecycle side by
side so admins see "approved at review, now archived" at a glance
URL consolidation:
* /store/{id} deleted (no redirect, stale bookmarks 404)
* /marketplace/flea/{id} is the canonical detail surface
* Three in-tree callers (upload-success, my-stack card, store
listing card) updated to point at the new URL
* Quarantine banner extracted to _quarantine_banner.html partial,
self-guarded, included from both flea detail templates
* Banner JS auto-refreshes when the verdict lands by polling
/api/marketplace/flea/{id}/detail (visibility_status +
submission_status — the latter is needed because blocked_llm
keeps the entity at visibility_status='pending')
Audit log resource format:
* runner.py emits prefixed `store_submission:{id}` (post-fix)
* Detail-page timeline query handles three patterns: prefixed
submission, helper-emitted `store_entity:{sub_id}`, and bare-id
legacy rows — all surface in the activity timeline
UX fixes:
* Owner sees Under review / Quarantined / Hidden banner with status
* Install button gray-disabled (not blue) when non-approved
* Owner cannot delete quarantined entries (403); admin can
* Admin queue: filter chips, sortable columns, paging, page-size
* Auto-refresh queue every 5s while pending rows are visible
* Store upload page file picker no longer opens twice (label →
input default action collided with explicit JS handler)
Tests: 168 passed across the guardrails suites (admin submissions,
store API, inline / LLM / purge guardrails, store repositories,
marketplace filter, schema version). New regression coverage
includes: archive surfaces via JOIN even when API path is bypassed;
deleted submission renders activity timeline (tombstone); flea
detail surfaces submission_status only for owner/admin; detail page
renders Entity lifecycle row; audit log resource format covers both
helper and runner paths.
* fix(store-guardrails): PR #233 follow-up — prompt injection, atomic PUT, BG race, schema, reaper, sort whitelist
Addresses 9 of the 23 findings from the PR #233 review (spec at
docs/superpowers/specs/2026-05-09-pr233-guardrails-fixes-spec.md).
Merge-gate items #1-#6 plus high-value mediums #7, #9-#12, #23.
Architectural items (#8 enum split, #14 factory) and pure
maintainability (#15-#22) deferred to follow-ups.
Security:
* #1 prompt injection — SYSTEM_PROMPT now passed via the SDK's
dedicated system= parameter; bundle wrapped in <bundle>...</bundle>
sentinels declared data-only by the system prompt; literal
sentinel strings in user content are escaped so an adversarial
README can't forge a close tag.
* #6 static scan honesty — module docstring + admin copy + docs
declare static scan as signal not gate; .md/.txt/.rst/.html/.json/
.yaml/.yml/.toml skipped to avoid false positives on prose.
AST mode for Python deferred (separate flag, FP comparison work).
Correctness:
* #2 PUT atomicity — bundles bake into plugin.staging-<rand>/
alongside live, atomic-rename on success; failed checks leave
live tree byte-for-byte intact.
* #3 BG-task race — set_visibility_if_pending guards verdict flips
to the (pending, hidden) review window; admin archives during
review survive; skipped flips audit-logged.
* #4 v35 NOT NULL/DEFAULT — schema v35→v36 re-applies them on
store_entities.visibility_status. CHECK constraint enforced
application-side (DuckDB ADD CHECK on existing column unsupported).
* #7 stuck-review reaper — reap_stuck_llm_reviews flips pending_llm
rows older than guardrails.stuck_review_grace_seconds (default
1800) to review_error. Scheduler runs every 15 min via new
/api/admin/run-reap-stuck-reviews. Set knob to 0 to disable.
* #9 quota counter — count_blocked_for_submitter_since now counts
blocked_inline + blocked_llm + review_error so a submitter
triggering only LLM-blocked verdicts is bounded.
* #10 missing risk_level — surfaces as review_error with
error='missing_risk_level' instead of silently defaulting to
'medium' (which looked like a model-decided block).
* #11 archived_at clear — set_visibility nulls archived_at +
archived_by when transitioning out of 'archived' so a future
read doesn't show stale archive forensics on an approved row.
Maintainability:
* #12 FSM doc comment — accurate insert/transition/lifecycle
description in src/db.py near store_submissions schema.
* #23 sort-key whitelist — admin queue rejects unknown sort keys
with 400 invalid_sort_key; substring-replace footgun removed.
Deferred (separate PRs):
* #5 quota race — proper fix requires asyncio.Lock spanning the
full pipeline; threading.Lock blocks event loop, DuckDB MVCC
doesn't help. API-level slowapi bounds worst case for now.
* #6 part 3 (AST static scan), #8 (enum split), #13 (import
bundle docs), #14 (factory consolidation), #15-#22 (maint).
Tests:
* New: tests/test_store_guardrails_prompt_injection.py (corpus +
trust-boundary invariants), tests/test_store_put_atomic.py,
tests/test_store_guardrails_reaper.py.
* Extended: test_store_guardrails_llm.py (system param, missing
risk_level, BG race), test_admin_store_submissions.py (quota
counter widening, sort whitelist 400), test_store_repositories.py
(un-archive metadata clear), test_db_schema_version.py (v36).
* Full suite: 3738 passed; 17 pre-existing baseline failures
unchanged (db migration tests, cli binary rename, catalog export,
user mgmt v5 backfill — confirmed by stash + rerun on clean tree).
* Extract session pipeline framework, refactor verification, add UsageProcessor skeleton
Pluggable framework under services/session_pipeline/ (contract + lib + per-processor
runner) so multiple processors can read /data/user_sessions/<key>/*.jsonl on their
own cadence with full failure isolation. Verification flow becomes the first plugin;
a no-op UsageProcessor reserves the second slot pending a separate brainstorm on
extraction logic + storage shape.
Schema v28→v29: rename session_extraction_state → session_processor_state with
composite PK (processor_name, session_file). Existing rows copied over with
processor_name='verification'; legacy table dropped. Migration is idempotent and
no-ops the copy step on fresh installs that came up at the new schema.
Endpoint: /api/admin/run-verification-detector replaced by parametrized
/api/admin/run-session-processor?processor=<name>. Audit action format follows.
Scheduler JOBS: verification-detector entry split into session-processor:verification
+ session-processor:usage. SCHEDULER_VERIFICATION_DETECTOR_INTERVAL retained for
operator compatibility (drives both cadence and health-check grace window);
SCHEDULER_USAGE_PROCESSOR_INTERVAL added.
* Address PR #232 review: scan dead branch + per-processor lock
- `SessionProcessorStateRepository.scan_unprocessed_for` dead else: both
branches surfaced every jsonl, the SELECT was unused, runner MD5-rehashed
every stable session per tick. Replaced with an mtime precheck — stable
sessions (mtime <= processed_at) are filtered at scan; modified files
still surface for the runner's authoritative `file_hash` invalidation.
Naive-local comparison matches the existing health-check idiom (DuckDB
TIMESTAMP strips tz on storage).
- Per-processor advisory lock around `_run_processor` in
`/api/admin/run-session-processor`. Scheduler tick + manual admin POST
could otherwise both run, both call create_evidence on overlapping
detections, and accumulate duplicate verification_evidence rows (the
dedup short-circuit only covers create+contradiction, not evidence per
ADR Decision 3). Non-blocking acquire → 409 Conflict on concurrent
invocation; release in finally so a runner exception doesn't wedge the
processor.
Tests: two new scan unit tests (mtime filter + post-mark mtime bump), 409
endpoint test, lock-released-on-exception test. Two existing tests updated
for the new "filtered at scan" stat shape (previously asserted skipped == 1,
now scanned == 0).
* Address PR #232 review #2: parallel scheduler tick + last_run on terminal state
Two pre-existing scaffold bugs in services/scheduler/__main__.py amplified
by adding more session-pipeline jobs:
1. Serial for-loop over jobs with synchronous httpx.post(timeout=900) — a
10-minute verification run blocked every other job (data-refresh,
health-check, usage, corporate-memory) for the whole window. The PR's
stated isolation guarantee held inside the runner but broke at the
scheduler dispatch layer.
2. last_run advanced only when _call_api returned True. Permanent-failure
jobs hot-looped on every tick (30s) instead of cadence (15min).
Fix: ThreadPoolExecutor.submit per due job + per-job in_flight set so a
long-running job can't be re-launched on subsequent ticks. last_run
advances unconditionally in finally; errors still surface via _call_api
logging + audit_log on the receiving side.
_run_job extracted to module-level for unit testing. New tests:
- TestRunJobBookkeeping: advances on success / failure / unhandled raise
- TestRunLoopParallelism: in_flight protection prevents duplicate
launches across ticks for a single slow job
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
* feat(home+news): state-aware /home + /news + admin-edited news section
Squash of the vr/home-page feature work for clean rebase onto main.
Original 18-commit history preserved in branch backup/vr-home-page-pre-rebase.
What's in this PR:
**State-aware /home page**
- New `/home` route with hero + auto-mode + connectors (Asana / GWS /
Atlassian) + lookarounds. Onboarded vs not-onboarded state-machine
branches a single template (`home_not_onboarded.html`); the install
steps, "Setup a new Claude Code" CTA (90-day PAT mint), and per-
connector setup prompts hide once `users.onboarded=TRUE`. A
completion badge replaces them.
- "Mark me as offboarded" button reverses the flag without an SQL UPDATE.
- `users.onboarded BOOLEAN` column added; default FALSE; flipped by the
CLI's `agnes init` post-success POST and the `/admin/users` API.
- Connector setup prompts pre-check whether the tool is already
installed/connected before re-running setup.
- GWS scope set widened to include Google Chat (`chat.spaces`,
`chat.messages`).
**Single template + design tokens**
- `dashboard.html` now extends `base.html` via the new
`{% block layout %}` opt-out (full-width pages skip the 800px
`.container`). Net: every page shares one shell.
- `style-custom.css` `:root` extended with `--space-{7,9,10,12}`,
`--radius-2xl`, `--shadow-{card,elevated}`, `--text-{muted,disabled}`,
`--focus-ring`, `--transition-*`, `--width-{narrow,app,wide}` so
inline page styles can migrate incrementally.
**Auth redirects honor AGNES_HOME_ROUTE**
- `safe_next_path` resolves the configured home route when no `default=`
is passed; OAuth callbacks, magic-link clicks, password form, and
LOCAL_DEV_MODE shortcuts now land on `/home` (or whatever the operator
picked) instead of always /dashboard.
**News section + /news permalink + /admin/news editor**
- Schema-bumped `news_template` table (single versioned entity, draft +
publish gate). `published BOOLEAN` distinguishes draft from public;
monotonically-increasing `version` per save; rows >30d pruned on
save except the currently-displayed published version.
- `/home` bottom-of-page renders the latest published intro with a
"Read more →" link to `/news` (which renders the full body).
- `/admin/news` editor with sandboxed live preview, versions table,
per-row Unpublish, Format-help cheatsheet.
- `agnes admin news show / draft / edit / publish / unpublish /
versions / export` (CLI). Talks to the live server via the
`/api/admin/news/*` endpoints (PAT-authed) — no direct DB access
so it coexists with a running uvicorn.
- **Optimistic-lock guard**: `agnes admin news publish --version N` and
PUT/PATCH endpoints accept `expected_version` and 409 with structured
`{error: "version_conflict", expected, actual, actual_by}` when a
concurrent admin replaced the draft. Edit refuses to overwrite a
draft authored by someone else without `--force` or
`--expect-version`.
- nh3 (Rust-backed ammonia) HTML sanitizer; iframe pre-pass strips
any iframe whose src is not on the YouTube/Vimeo/Loom allowlist;
javascript:/data: schemes blocked everywhere.
- Author CSS vocabulary: `.news-hero` (blue gradient hero block),
`.callout`/`.callout-{info,warn,success,danger}`,
`.video-embed`, `.news-section`, `.news-grid-{2,3}`, `.news-cta` —
all consolidated in `style-custom.css` under "News content
vocabulary (shared)" so /home perex, /news body, and /admin/news
preview share one source of styling.
- Code-inside-`<pre>` contrast fix (was unreadable amber-on-silver).
- `.news-content` table styling (border, header band, row-hover).
**`scripts/dev/run-local.sh`** — local uvicorn launcher. Pulls Google
OAuth client id/secret from GCP Secret Manager
(`AGNES_OAUTH_GCP_PROJECT`-driven, no vendor defaults), points
`AGNES_CLI_DIST_DIR` at `./dist` so the wheel endpoint resolves, and
`--dev` flips `LOCAL_DEV_MODE=1` + `AGNES_HOME_ROUTE=/home` for one-
command iteration. `LOCAL_DEV_MODE=1` also enables the FastAPI debug
toolbar.
**CLAUDE.md "Run tests before every push" section** codifies
`pytest tests/ -n auto -q` as non-negotiable before each push.
**Tests**: 51 + 14 + 8 = 73 new tests across news-template repo,
sanitizer, API, web, CLI; plus updated home/auth/template tests for
the new shared-shell architecture.
Origin docs (gitignored, customer-fork content):
docs/brainstorms/home-page-requirements.md,
docs/plans/2026-05-07-001-feat-home-page-plan.md.
* feat(cli): agnes onboarded {on,off,status} — self-scoped flag toggle
User-facing equivalent of the in-page "Mark me as (off)boarded" button
on /home. POSTs /api/me/onboarded with {onboarded, source}; --source
overrides the audit-log marker so flips made from the CLI vs the web
button vs agnes init automation stay distinguishable.
`status` reads via /api/me/profile (when present); falls back to a
quick body-marker scan of /home so the read path doesn't write an
audit_log row. PAT-authed via cli.client.api_post — same convention
as agnes admin news / agnes admin add-user etc.
Tests: 5 covering on/off/status round-trip, idempotency, and
audit-log source recording. Full suite holds at 12 pre-existing
failures (same set as before).
* ui(nav+home): primary nav reorg + green What's new band + /marketplace link fix
Primary nav (post-rebase audit + per-user feedback):
- Items: Home → Marketplace → Data Packages → Memory. Admin dropdown
for admins only. The "Dashboard" label was renamed Home — point still
resolves through `home_route` so customer instances on /dashboard
still land there.
- Activity Center moved into the Admin dropdown. Per-team adoption
analytics is admin-consumed in practice; the route still allows
any authed user for direct deep-links so existing /home tile +
bookmarks keep working.
- Memory link added (→ /corporate-memory) — was previously buried in
the /home "Look around" tiles.
- Setup local agent + My Stack dropped from main nav. Setup is the
/home install flow's home now; My Stack lives as a tab inside
/marketplace.
/home tweaks:
- Plugin marketplace tile now points at /marketplace (was /store —
legacy from before the marketplace rebrand landed in #230).
- "What's new" section header gets a green band (success-flavored
D1FAE5 background, A7F3D0 border, darker green title) so the
bottom-of-page news block visibly distinguishes from the blue
install-hero at the top. Header strip only — body stays white.
Test fix: test_home_route_resolution renamed `dashboard_link_uses_home_route`
→ `home_link_uses_home_route` and asserts `href="/home">Home` instead
of `href="/home">Dashboard` after the label change.
* fix(home): decouple Step 3 + Connect-tools collapse from server onboarded flag
The server-side `users.onboarded` flip happens through two paths:
1. Explicit user click on "Mark me as onboarded" or `agnes onboarded on`.
2. Implicit `agnes init` POST → /api/me/onboarded on success.
Path 2 produced a UX surprise: an analyst running `agnes init` mid-flow
reloaded /home and saw Step 3 (auto-mode) + Connect-your-tools auto-
collapse to summary bars. They were actively working through those
sections — the install POST never signalled "I'm done with the rest
of setup", just "Agnes itself is installed".
Decouple the section-collapse decision from the server flag:
- Step 1 + Step 2 install blocks: still hidden on `onboarded=TRUE`
(their completion is a hard server signal — Agnes IS installed).
- Step 3 + Connect-your-tools: render flat by default in BOTH states.
Wrapped in `<details class="setup-collapsible" open>` so the
browser's native disclosure handles per-section toggle without JS,
but the `<summary>` is CSS-hidden until the page-level
`data-setup-minimized="1"` attribute is set on `.home-mock`.
- New "Minimize setup view" toggle inside the blue install-hero,
rendered only when onboarded. Click flips the data-attr on
`.home-mock` AND removes the `open` attribute from each
`<details>`. State persists in `localStorage["agnes_home_setup_minimized"]`
so the choice survives reloads but is per-device.
- "Show full setup view" (the same button when minimized) re-opens
both `<details>` and clears localStorage.
When minimized, each `<details>` still has its own native expand/
collapse — click the gray summary bar to peek at one section without
toggling the page-level minimize off.
Tests:
- test_step3_and_connectors_render_flat_when_onboarded_by_default —
asserts `<details class="setup-collapsible" ... open>` for both
sections post-onboarding and the absence of any server-rendered
`data-setup-minimized` attribute on the `.home-mock` root.
- test_minimize_toggle_visible_only_when_onboarded — toggle button
rendered only when onboarded.
Full pytest holds at 12 pre-existing failures (same set).
* feat(observability): optional PostHog integration (errors, LLM traces, replay, flags)
Off by default. Activates when POSTHOG_API_KEY is set in env. Defaults
to PostHog Cloud EU; override host for US Cloud or self-hosted.
Coverage:
- FastAPI 500 handler captures unhandled exceptions
- src/orchestrator.py rebuild + rebuild_source failures
- services/scheduler/ HTTP-job failures
- cli/main.py uncaught CLI errors (Typer.Exit/SystemExit/KeyboardInterrupt
skipped; flushes before re-raise so short-lived CLI invocations don't
drop events)
- connectors/llm/anthropic_provider.py + openai_compat.py emit
$ai_generation events with provider, model, latency, token counts
(prompt/completion bodies stay off unless POSTHOG_LLM_PAYLOADS=1
because LLM prompts here routinely include customer SQL/data)
- Browser snippet injected into every text/html response by
PosthogInjectionMiddleware — registered inside the GZip layer so it
sees uncompressed HTML before compression. Many templates are
standalone (their own DOCTYPE) and never extend base.html, so a
per-template include would miss them.
- Frontend: $pageview, $pageleave, JS error capture via window.error
and unhandledrejection handlers, masked session replay
(maskAllInputs: true plus CSS-selector mask for known data surfaces),
feature flags (browser posthog.isFeatureEnabled + server-side
feature_enabled with fallback for older SDKs).
Identification mode operator-configurable: none / id / email / full.
Default email ships user.id + email but never name. CLI entry point
moves from cli.main:app to cli.main:main (Typer wrapper).
Files:
- src/observability/posthog_client.py — lazy singleton, no network
when disabled, single-process flush on shutdown
- src/observability/llm_tracing.py — trace_generation context manager
- app/middleware/posthog_inject.py — HTML rewrite middleware
- app/web/templates/_posthog.html — browser snippet template
- docs/observability.md — operator guide
- config/.env.template — documented POSTHOG_* knobs
- tests/test_posthog_disabled.py + tests/test_posthog_client.py +
tests/test_llm_tracing.py — 18 tests covering disabled state,
identify-mode payloads, $ai_generation shape, error variant.
CHANGELOG entry under [Unreleased] Added.
* feat(observability): tag every PostHog event with environment + release
Splits PostHog dashboards cleanly between localhost / dev / staging /
production without manual tagging on every capture call.
- POSTHOG_ENVIRONMENT explicit override; auto-resolves to "local" when
LOCAL_DEV_MODE=1, else RELEASE_CHANNEL, else AGNES_DEPLOYMENT_ENV,
else "unknown".
- AGNES_VERSION → RELEASE_CHANNEL fallback feeds the `release` property
for "is this error new in this release?" cohorting.
- Backend gets both via the PostHog SDK's super_properties constructor
arg (every captured event picks them up automatically).
- Browser snippet calls posthog.register({environment, release}) inside
the loaded callback so $pageview, $exception, autocapture, etc. all
carry the same labels.
- request.state.user now populated by auth dependencies so the snippet
can actually call posthog.identify(user_id, {email}) for logged-in
users (previously the user block always resolved to None because
nothing wrote to request.state.user).
4 new tests cover env resolution: explicit > LOCAL_DEV_MODE > channel
> unknown, plus super-properties forwarding into the SDK constructor.
* feat(observability): inline user attrs on every PostHog event + debug throw route
PostHog's UI shows person properties on the Person profile page, not
inline on each event — so a reviewer triaging an exception couldn't tell
which user hit the bug without clicking through. Fix it on both sides.
- Backend capture_exception merges user_id / user_email / user_name into
the event properties (gated by POSTHOG_IDENTIFY_PII: none/id/email/full).
Backed by a new _user_props_for_event helper on PosthogClient.
- Browser snippet registers user_id + user_email + user_name as super-
properties via posthog.register({...}) so every $exception, $pageview,
and custom event coming from posthog.captureException() carries them
inline. Mirrors the backend so cross-referencing client/server events
doesn't require a person-profile lookup.
- /api/debug/throw — debug-only endpoint gated by DEBUG=1 (404 in prod).
Runs Depends(get_current_user) first so request.state.user is set when
the unhandled-exception handler captures the event. Lets operators
exercise the full observability path end-to-end without hand-rolling
a TestClient script. Configurable via ?kind=ValueError&msg=...
7 new tests cover: backend user-attr merge across identify modes,
anonymous request fall-through, browser snippet super-prop emission for
logged-in / anonymous / id-only / full-name cases.
* fix(observability): address minasarustamyan PR #231 review
Two bugs caught in review.
1. PosthogInjectionMiddleware dropped Response.background on every
return path. BaseHTTPMiddleware materialises the body and asks
subclasses to return a fresh Response — three paths in dispatch()
omitted background=, silently cancelling any BackgroundTask /
BackgroundTasks the route attached (audit logging, async webhooks,
email sends) with no log line. Fix: route every return through a
_passthrough() helper that forwards background.
Also adds a _MAX_BUFFER_BYTES (4 MB) cap so a streamed-HTML response
can't balloon RSS during buffering. Bigger bodies short-circuit
through with a warning rather than being injected.
Regression tests in tests/test_posthog_inject_middleware.py exercise
four return paths (snippet present, render-fail, double-injection
guard, non-HTML passthrough) plus the streaming-guard short-circuit.
2. $ai_input / $ai_output_choices were emitted without truncation, so
POSTHOG_LLM_PAYLOADS=1 silently dropped events past PostHog's ~32 KB
per-event ingest limit — exactly the calls (large prompts with
schemas / sample rows / SQL) an operator would want to inspect.
Fix: clip both at POSTHOG_LLM_PAYLOAD_MAX_CHARS (default 30000) with
an explicit "…[truncated N chars]" marker so readers don't mistake
truncated captures for complete ones. Metadata (provider, model,
tokens, latency, error) flows regardless. Three new tests cover
default-cap clipping, env-override, and pass-through under the cap.
37 PostHog tests pass.
* Add /marketplace browse page + Model B opt-in stack composition
New /marketplace browse surface unifies the curated marketplaces
(admin-managed git mirrors) and the community Flea Market behind
three tabs — Curated / Flea / My Stack — with per-tab category
filter, search across both sources with scope checkboxes, and
numeric pagination, all driven by URL query state. Plugin detail
at /marketplace/curated/<slug>/<plugin> and /marketplace/flea/<id>;
nested skill / agent detail at /marketplace/curated/<slug>/<plugin>/
{skill,agent}/<name> and the flea-side single-page detail.
Model B opt-in: an RBAC grant on a curated plugin is now only
*eligibility*. The user must click "Add to my stack" for it to
enter their served Claude Code marketplace. Composition flips
from (rbac ∖ opt_outs) ∪ store_installs to
(rbac ∩ subscriptions) ∪ store_installs. The legacy
user_plugin_optouts table is renamed user_curated_subscriptions
(schema v27) — same table shape, inverted semantic, repository
methods become subscribe / unsubscribe / is_subscribed.
UX vocabulary: Install → Add to my stack, Installed → In your
stack, card "Installed" badge → "In stack" (amber pill), tab
"My Subscriptions" → "My Stack". Bridges the two-step model
(server-side bookmark vs. on-laptop install) the previous label
hid. Click triggers an inline post-add hint panel under the
description with the agnes refresh-marketplace recipe + Copy
chip, dismissible per-browser via localStorage.
Per-tab info blocks above the filter row:
- Curated: trust signal — "Each plugin here has a named curator
accountable for it." (blue accent + See-all-curators link)
- Flea: open-shelf signal — "Anyone in the company can upload
here." (purple accent + Tips-for-sharing link)
- My Stack: personal-shelf orientation — "Your AI stack —
everything you've added." (slate accent, no link)
Tabs carry per-tab Heroicons (shield-check / building-storefront
/ rectangle-stack) tinted to match each tab's accent; flips white
when the tab is active for contrast.
Hero illustration anchored to the right of the blue hero panel
(absolute, 47% wide, behind the search row content). Hidden
under 900px viewport.
Action-row CTAs realigned to publication intent: curated
"How to add new content" → "Submit a plugin" (links to the
guide page); flea button removed since +Upload sits next to it.
Empty-state CTAs match. /marketplace/guide/{curated,flea}
routes now host publication-flow guide pages with placeholder
ledes — full copy to be authored separately.
Categories: Heroicons-based icons mapped per category in
src/category_icons.py (zero new dependencies; SVG path strings
inlined). Marketplace cards, filter pills, and detail pages
read from the same source.
API endpoints under /api/marketplace:
- GET /items per-tab listing (curated / flea / my)
- GET /categories per-tab non-zero counts
- GET /curated/{slug}/{plugin} plugin detail
- POST/DELETE /curated/{slug}/{plugin}/install subscribe toggle
- GET /curated/{slug}/{plugin}/{skill,agent}/{name} inner item
The tab=my branch reads directly from
user_curated_subscriptions ∪ user_store_installs (not
resolve_user_marketplace, which bundles flea skills/agents into
a single store-bundle synthetic entry useful for serving the
Claude Code marketplace ZIP/git but wrong for browsing where
each item should appear as its own card).
Detail pages: plugin detail surfaces inner skills/agents as
clickable nested cards; commands/hooks/MCPs render as plain
name lists. Skill/agent detail mirrors the plugin layout with
kind-tinted accents (skill = green, agent = purple), Description
+ Details sidebar, Files + Docs sections, and the "How to call
it" copy-able invocation chip showing /<plugin>:<inner-name>
exactly as Claude Code namespaces it post-install. Curated
nested has no install button — links back to the parent plugin.
Navbar: standalone "My AI Stack" relabelled "My Stack" and
points at /marketplace?tab=my; "Store" link removed (Store
flow is reachable via the Flea Market tab's +Upload button).
The standalone /my-ai-stack and /store routes still work for
old bookmarks.
Tests cover the new browse / categories / install / RBAC paths
under tests/test_marketplace_api.py; existing marketplace and
store tests updated for Model B (explicit subscribe in fixtures).
Schema bumped v26 → v27 with idempotent migration that wipes
existing user_plugin_optouts rows on flip and adds
marketplace_plugins.created_at with registered_at backfill.
* Fix v28 migration + post-rebase test fallout
v28 ALTER TABLE marketplace_plugins ADD COLUMN created_at conflicted with
_SYSTEM_SCHEMA's earlier CREATE that already includes the column on fresh
installs (test fixtures starting at any pre-v28 version trip on it).
Switch to ADD COLUMN IF NOT EXISTS — same idiom as the upstream v27
Keboola sync-strategy migration on the same ladder.
Two test patches needed after the rebase bumped SCHEMA_VERSION 27 → 28:
- test_keboola_v27_migration.py: test_schema_version_constant_is_27 was
pinning ==27. Loosened to >=27 (the test's purpose is to verify the
v27 Keboola migration, not to pin the current SCHEMA_VERSION).
- test_setup_page_unified.py: was monkeypatching resolve_allowed_plugins
but compute_default_agent_prompt now reads from resolve_user_marketplace
(Model B-aware). Stub the right function so the test exercises the
v28 served-set path.
* Harden curated skill/agent inner endpoints against path traversal
`_read_inner`, the `skill_dir` walk in `curated_skill_detail`, and the
`agent_path.stat` in `curated_agent_detail` joined URL path-params onto
`plugin_root` without verifying the resolved candidate stayed inside it.
Starlette's `[^/]+` on `{skill_name}` / `{agent_name}` blocks the direct
URL exploit (encoded `/` 404s before the handler), but a curator-planted
symlink inside a curated marketplace's git mirror could still dereference
outside the plugin tree on read.
Adds `_safe_join(plugin_root, *parts)` doing
`Path.resolve(strict=True)` + `relative_to(plugin_root.resolve())`, used
by all three call sites so the boundary is enforced once and consistently.
Tests cover the helper directly (normal path resolves, escaping `..`
returns None, escaping symlink returns None, missing file returns None)
plus an end-to-end check that the symlink case actually 404s on the
HTTP endpoint. Symlink tests skip on Windows where symlink creation
needs elevated permissions; they run on Linux CI.
---------
Co-authored-by: Minas Arustamyan <arustamyan.minas@gmail.com>
## Summary
Two minimum-viable fixes after today's 0.44.0 → 0.47.3 release train and the production 30-user launch. Devil's advocate review of a 3-PR / 7-item plan cut scope to these 2 — the rest is deferred to a separate "operate-first, instrument-second" backlog item.
### B2 — Docker session_collector log skip
`services/session_collector` was logging `Collection complete: 0 users, 0 files copied` + `WARNING: Group 'data-ops' not found, using default group` every 10 minutes in the Docker layout (where `/home/*/user/sessions/` doesn't exist). New env var `AGNES_SKIP_LEGACY_COLLECTOR=1` set by default in `docker-compose.yml` short-circuits the collector pass.
The bare-VM deployment path (where /home/* IS populated by Claude Code) leaves the env var unset and continues to scan normally — including the data-ops warning, which is load-bearing for catching missing-group mis-deploys.
### O2 — FIFO check in `_check_session_pipeline`
The existing check compares `MAX(processed_at)` to newest jsonl mtime — catches "detector hasn't run lately" but blind to "old file was skipped while newer ones were processed". New code finds the oldest FS jsonl that's NOT in `session_extraction_state.session_file` and flags if its mtime is older than `SESSION_PIPELINE_STUCK_FILE_GRACE_SECONDS` (default 4× the existing grace = 2h).
Severity intentionally starts at `info` so we can collect prod data on false-positive rate before tightening to `warning`. The aggregator already treats `info` as non-promoting (see the severity vocabulary docstring at the top of `app/api/health.py`), so the headline `status` stays at `healthy` even when this fires — the operator sees the entry in the per-check breakdown but no spurious `degraded` overall.
## Test plan
- [x] `pytest tests/test_session_collector.py` — 17 tests pass (existing 9 + new 8 covering env-set/unset, truthy variants, falsy non-skip).
- [x] `pytest tests/test_health_session_pipeline.py` — 8 tests pass (existing 4 + new 4 FIFO tests covering stuck-file, under-threshold, all-processed, env-override).
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/229" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
`agnes self-upgrade` without `--force` previously short-circuited on the local 24h `update_check.json` cache. After a server-side version bump within that window, the explicit command exited silently as a no-op — empirically observed today when prod 0.47.1 → 0.47.2 didn't propagate.
Fix: always invalidate the cache in `_resolve_info`. The cache still gates the implicit warning loop in the root callback (correctly — that runs on every `agnes <anything>` and can't hammer `/cli/latest`).
## Test plan
- [x] New `test_self_upgrade_bypasses_24h_cache_without_force` — stale cache claims current; mocked server reports newer; assert UpdateInfo carries the newer version, not the cached one.
- [x] Existing self-upgrade tests pass (including `--force` semantics — force is now downstream-only, behavior preserved).
## Summary
Smoke-testing the just-shipped 0.47.1 against production exposed two regressions:
1. `agnes query --remote "SELECT FROM unit_economics WHERE bad_col=1"` returned `Table "unit_economics" must be qualified` (the OLD error) instead of `Unrecognized name: bad_col` (the #218 fix's intended behavior).
2. `agnes query "DESCRIBE unit_economics"` showed only DuckDB's misleading `Did you mean order_economics?` with no Agnes hint paragraph (the #219 fix is missing).
Root cause: PR #217's squash merge (`506a378c`) carried stale snapshots of `app/api/query.py` and `cli/commands/query.py` from before #218 and #219 merged. The rebase-and-merge auto-merged those files cleanly (no conflict markers) but the result silently reverted both fixes.
Restore the two changes verbatim. Tests for both fixes already on main and continue to pass against the restored code.
## Test plan
- [x] `pytest tests/test_api_query_guardrail.py tests/test_cli_query.py` — clean
- [x] Manual repro against prod after deploy: both flows now surface the intended diagnostic.
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/225" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
Brings the Keboola connector to feature parity with the legacy internal data-analyst's per-table sync strategies. Closes the four documented gaps from the spec branch (`zs/keboola-connector-specs`):
- **Typed parquet** in the legacy SDK extraction path — column types from Keboola Storage metadata (provider cascade `user > ai-metadata-enrichment > keboola.snowflake-transformation`) survive the CSV → parquet roundtrip; invalid date strings (`'0000-00-00'`) and invalid numeric strings (`'Non-Manager'`) become NULL while keeping the column's typed schema. Pre-fix everything was VARCHAR.
- **Incremental sync** via Storage API `changedSince` — opt-in per table; pulls only delta rows, merges into the existing parquet by `primary_key` (drop_duplicates with keep='last'). Cuts daily extraction from O(full table) to O(delta).
- **Partitioned sync** — flat per-partition layout `data/<table>/<key>.parquet` (e.g. `2026_05.parquet`), per-affected-partition merge for daily updates, chunked initial load with 1-day overlap and 2-empty-chunk stop heuristic.
- **`where_filters`** — server-side row filter with date placeholders (`{{today}}`, `{{last_3_months}}`, `{{start_of_3_months_ago}}`, etc.) resolved at sync time. Force the SDK path; reject `incremental + where_filters` combination at API layer (changedSince already filters temporally).
## Architecture
- **Schema migration v25 → v26**: 7 new columns on `table_registry`. Existing `sync_strategy` column reused (pre-v26 it was inert catalog metadata; post-v26 the extractor dispatches off it).
- **Per-table dispatcher** in `extractor.run()` routes to one of `_extract_via_extension` (full_refresh + extension), `_extract_via_legacy` (full_refresh + filters or extension fallback), `extract_incremental`, or `extract_partitioned`.
- **API conflict policy**: `incremental + where_filters` → 422; `partitioned + query_mode='remote'` → 422; `partitioned ⇒ partition_by required`.
- **Admin UI**: third "Direct extract (Storage API)" radio in the Keboola Register / Edit modals, alongside existing "Whole table (extension)" and "Custom SQL". When selected, exposes a v26 sync-strategy panel with conditional fields per strategy.
## Test plan
- [x] **Unit + module** — 134 v26 tests covering migration, repo, parquet_io, where_filters, incremental (compute_changed_since + merge_parquet + extract_incremental E2E), partitioned (key derivation + merge_partition + chunked windows + extract_partitioned E2E), extractor dispatcher, admin API validators, PUT field clearing, registry-shape → dispatcher bridge
- [x] **HTML form structure** — all v26 inputs + visibility classes + JS payload fields verified in rendered template
- [x] **Real Keboola roundtrip** — registered a small test table as `sync_strategy='incremental'` against a test Storage project, triggered two syncs:
- Sync 1: `changedSince=None` → full pull → 9 rows typed parquet
- Sync 2: `changedSince=last_sync - 1d window` → 9 delta rows merged with 9 existing → 9 after dedup on primary_key (PK merge confirmed)
- [x] **Browser UX** — agent-browser session against a local uvicorn: login → admin/tables → register modal → switch radios → verify field visibility per strategy → submit → edit existing row → switch to Direct/Incremental → save → confirm DB persistence
- [x] **Regression** — no regressions in the broader 3252-test suite (3 pre-v26 tests updated for the deprecation-marker removal + schema-version bump; 2 pre-existing environment-sensitive test failures unrelated to this change)
## Bugs caught + fixed during E2E
The browser + real-Keboola roundtrip exposed four bugs the unit tests missed:
1. **JS visibility race** — two competing `forEach` loops set `display=''` then `display='none'` on form elements sharing `kb-strategy-incremental kb-strategy-partitioned` classes (window_days + max_history_days are reused across strategies). Fix: single-pass selector with class-based visibility resolver.
2. **PUT cannot clear field** — pre-v26 `updates = {k: v ... if v is not None}` collapsed "omitted from body" and "sent as null" into the same case, so admin couldn't switch a partitioned row back to full_refresh and have stale `partition_by` clear. Fix: `model_dump(exclude_unset=True)`.
3. **Subprocess DB lock conflict** — `_read_last_sync` reopened `system.duckdb` while the parent server held the write lock (subprocess contract at `app/api/sync.py:_run_sync` line 260). Fix: parent injects `__last_sync__` into table_config before subprocess spawn.
4. **Wrong KBC table_id** — `extract_incremental` / `extract_partitioned` built the Storage API table_id from the registry row's slugified `id` (`circle_inc`) instead of `bucket.source_table` (`in.c-finance.circle`), producing 404s. Fix: prefer `bucket+source_table`; fall back to `id` only when bucket empty.
## Operator notes
- Existing tables stay on `full_refresh` after migration; admins opt individual tables in via `agnes admin register-table --sync-strategy ...`, the Keboola Edit modal, or `POST/PUT /api/admin/registry`.
- `merge_parquet` and `merge_partition` use `pd.concat + drop_duplicates`, loading both existing and delta into pandas RAM. For tables in the multi-million-row range this may OOM — switch to `partitioned` strategy for those (per-partition merge keeps memory bounded). Documented in `### Internal` of the changelog entry.
- Date placeholders are resolved at **sync time**, not register time — a typo'd `{{lasst_week}}` is accepted at register and surfaces only when the next sync runs. By design (rolling windows need late-binding).
## Spec source
The four corresponding plans on the `zs/keboola-connector-specs` branch under `docs/superpowers/plans/2026-05-07-0[1-4]-*.md` capture the design rationale and link back to internal repo references for each subsystem.
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/217" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
- Catalog enrichment for `query_mode='remote'` rows: `rows`, `size_bytes`, `partition_by`, `clustered_by` per table (BQ + Keboola providers).
- `/api/v2/schema/{id}` cache miss: 2 BQ jobs → 1 (-50%) via shared `fetch_bq_columns_full`.
- All four catalog/schema/sample/metadata caches flush on registry change; single-row re-warm scheduled.
- Automatic cache warmup at server startup (bounded concurrency, opt-out via `AGNES_SKIP_CACHE_WARMUP=1`).
- SSE-driven freshness toolbar on `/admin/tables` with progress bar, log, and per-row badge.
- New admin doc `docs/admin/query-modes.md` — single source of truth on `local` / `remote` / `materialized` choice.
Closes#155.
Closes#156.
## Test plan
- [x] 65+ targeted tests pass across 11 new test modules + 3 modified ones.
- [x] No DB migration; no wire-break; `MIN_COMPAT_CLI_VERSION` unchanged.
- [ ] Reviewer: register a remote BQ table via `/admin/tables`, observe the toolbar populates within ~2 s and the per-row badge transitions warming → fresh.
- [ ] Reviewer: trigger `Re-warm all`, verify SSE log scrolls and `cacheWarmupBar` progresses.
- [ ] Reviewer: edit a registered row's bucket, verify `agnes schema <id>` returns updated columns immediately (no 1-hour staleness).
- [ ] Reviewer: confirm `agnes admin register-table --query-mode remote` prints the new IAM-smoke-check hint.
## Notable design decisions
- BigQuery `INFORMATION_SCHEMA.TABLE_STORAGE` is the only valid scope for size+rows (verified live 2026-05-07; dataset-scoped doesn't exist). Region resolved from `instance.yaml.data_source.bigquery.location` → `bq.client().get_dataset(...)` → fall back to legacy `__TABLES__`.
- VIEW handling: TABLE_STORAGE returns no rows for views, fall through to `__TABLES__` (also empty) → `TableMetadata(rows=None, size_bytes=None, partition_by=..., clustered_by=...)`. Null size signals analyst Claude to apply existing CLAUDE.md guidance.
- `size_bytes` is `active_logical_bytes + long_term_logical_bytes` — full BQ scan reads both; reporting only active undercounts aged partitioned tables.
- Source-agnostic provider seam: per-source `connectors/<source>/metadata.py:fetch(MetadataRequest)`; dispatcher in `app/api/v2_catalog.py:_metadata_provider_for` lazily imports per source_type so a Keboola-only deployment doesn't pay the BQ-extension import cost.
- Warmup non-blocking: FastAPI `lifespan` schedules `asyncio.create_task(_warm_catalog_caches_bg)` before `yield`. Per-row failures isolated.
## Out of scope
- Profile / column histograms / dimension cardinality for remote tables (separate issue).
- Onboarding nudge ("you have 0 remote tables, consider registering some BQ ones") — separate UX call.
- Provider plug-in registration via entry-points (the dispatch table is a hardcoded if-tree today; one line per future source).
## Release
Bumps `pyproject.toml` 0.46.1 → 0.47.0 (main shipped 0.46.0 + 0.46.1 during this PR — see commit `d98976ec`). New CHANGELOG section under `## [0.47.0] — 2026-05-07`.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/223" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
Two bugs in `agnes describe` surfaced from a real analyst session following the CLAUDE.md agent-rails discovery workflow. Together they break `agnes describe` end-to-end for any analyst (or analyst-AI) who follows the documented form.
### A) CLI parsing
`agnes describe TABLE -n 5` failed with `Missing argument 'TABLE_ID'`. Root cause: the command was registered as a `Typer.Typer` subcommand group via `app.add_typer(describe_app, name="describe")` + `@describe_app.callback(invoke_without_command=True)`, and that pattern mis-parses positional + short-int option in some orderings. Same pattern in `cli/commands/schema.py` works only because schema has no INTEGER short option. Fix: switch to flat `@app.command("describe")`.
### B) Server NaN
`/api/v2/sample/<id>` (called by `agnes describe`) returned HTTP 500 with `ValueError: Out of range float values are not JSON compliant: nan` whenever a row contained NaN. Fix: sanitize NaN/±inf to None before JSON serialization.
## Test plan
- [x] `pytest tests/test_cli_describe*.py` — added regression tests pinning `-n` parsing on either side of the positional.
- [x] `pytest tests/test_api_v2_sample*.py` — added regression test for NaN row → JSON `null` (not 500).
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/224" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
`claude -p` (headless mode) gives SessionEnd hook subprocesses ~1 second before SIGTERM, regardless of work in progress. `agnes push` for a typical workspace takes 5-30s. The current synchronous SessionEnd hook (`agnes push --quiet 2>/dev/null || true`) was therefore being killed mid-first-upload — `|| true` masks the SIGTERM as exit 0, so this regression was invisible until I traced it via a wrapper script and Claude's `~/.claude/debug/<sid>.txt` log.
Fix: wrap SessionEnd push in `bash -c "( nohup agnes push --quiet </dev/null >/dev/null 2>&1 & ) ; true"`. The subshell exits immediately, orphaning the upload child to init so it survives the hook subprocess kill. Same `bash -c` pattern as the existing `refresh-marketplace` SessionStart entry (for Windows compatibility).
End-to-end verified against production: claude exited in 5s, detached child completed the upload, file `491e3a23-...jsonl` landed on the server within 30s with mtime 14:30 UTC.
## Test plan
- [x] `pytest tests/test_lib_hooks.py` — added `test_session_end_push_is_detached` regression test asserting `nohup`, `&`, `</dev/null` are all present.
- [x] `pytest tests/test_setup_hooks_template.py` — assertions loosened from `==` to `in` where necessary.
- [x] Verified end-to-end against production with the detached wrapper before opening this PR (manual probe).
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/222" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
Verified against production: `claude -p` headless mode doesn't fire SessionEnd hooks (proven via `--output-format stream-json --include-hook-events`: zero `SessionEnd` events), so any session JSONLs from `-p` invocations stay orphaned locally and never reach the server. Fix: add `agnes push --quiet` as a third SessionStart entry — symmetric self-heal alongside the existing `agnes pull` entry. Existing workspaces pick this up on their next `agnes init` via the marker-based migration already in `cli/lib/hooks.py`.
Separately: a colleague's fresh install showed `agnes diagnose` warning "uploads are not being processed", which led them to suspect their `agnes push` was broken. The warning is actually about the LLM-based `verification-detector` backlog (uploads themselves were arriving fine — confirmed by 23+3 JSONLs landed on the server while the warning was firing). Reword the warning to "verification-detector backlog" + add `last_processed` to the diagnose dict so operators don't have to grep logs to confirm.
## Test plan
- [x] `pytest tests/test_lib_hooks.py` — updated count + added `agnes push in SessionStart` assertion.
- [x] `pytest tests/test_setup_hooks_template.py` — updated.
- [x] `pytest tests/test_clean_install_integration.py` — updated.
- [x] `pytest tests/test_health_session_pipeline.py` — updated warning text + asserted `last_processed` field.
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/220" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
`agnes query "DESCRIBE unit_economics"` (where `unit_economics` is `query_mode='remote'`) previously returned DuckDB's nearest-name suggestion (`Did you mean "order_economics"`?), sending users down the wrong path. Now appends a friendly hint about remote tables.
Reproduced from a real analyst session — colleague spent ~30s diagnosing what was actually "this is a remote table, not materialized locally".
## Test plan
- [x] New test: `_query_local("DESCRIBE unit_economics", ...)` against an empty local DuckDB triggers the new hint, original DuckDB error still echoed.
- [x] Negative test: a syntax-error query does NOT trigger the hint (regex only matches "Table with name X does not exist").
- [x] `pytest tests/test_cli_query*.py` clean.
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/219" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
## Summary
When `agnes query --remote` references a column that doesn't exist on the FROM table, users were seeing `Table "<id>" must be qualified with a dataset` instead of the actually-useful `Unrecognized name: <column>` from BigQuery. Surface the first-attempt diagnostic now; keep the second-attempt context as `underlying_original`.
Reproduced against production:
```
$ agnes query --remote "SELECT COUNT(*) FROM unit_economics WHERE authorize_date = DATE '2025-05-06'"
Error: remote_estimate_failed (HTTP 400)
message: Could not estimate scan size for this query.
underlying: 400 ... Table "unit_economics" must be qualified with a dataset.
```
(`unit_economics` has `authorize_timestamp`, not `authorize_date`.)
## Test plan
- [x] New `test_remote_estimate_failed_surfaces_first_error_when_attempts_differ` asserts the first-attempt message wins, second-attempt is preserved as `underlying_original`, hint points to `agnes schema`.
- [x] Existing `test_guardrail_returns_400_remote_estimate_failed_on_double_parse_error` still passes (both attempts mocked to identical error).
- [x] `pytest tests/test_api_query_guardrail.py` clean.
<!-- devin-review-badge-begin -->
---
<a href="https://app.devin.ai/review/keboola/agnes-the-ai-analyst/pull/218" target="_blank">
<picture>
<source media="(prefers-color-scheme: dark)" srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1" alt="Open in Devin Review">
</picture>
</a>
<!-- devin-review-badge-end -->
* fix: cutover regressions + parallel Keboola legacy fallback
Bundled fixes from a fresh-deploy run on a Keboola Storage backend with
the block-shared-snowflake-access feature flag — DuckDB Keboola
extension's per-table scan can't access bucket schemas, so the legacy
kbcstorage Storage-API client is the only working path.
CUTOVER REGRESSIONS
- agnes pull hash mismatch on every Keboola local-mode table —
src/orchestrator.py:_update_sync_state stored md5(mtime+size)[:12]
while the CLI compares against full 32-char content MD5. Now stores
the same content MD5 the materialized SQL path already used.
- Trailing-slash sanitization in connectors/keboola/access.py and
extractor.py — DuckDB Keboola extension's ATTACH fails when the URL
ends in / (canonical form).
- src/profiler.py:TableInfo.description becomes optional — two call
sites instantiated without it, crashing the profiler pass.
- scripts/ops/agnes-auto-upgrade.sh: chown on UID change — older images
ran as root, current runs as agnes (uid 999). Reads target uid:gid
from /etc/passwd inside the new image and chowns ${STATE_DIR},
/data/extracts, /data/analytics when the digest moves.
- POST /api/sync/trigger is now singleton per process — two
near-simultaneous trigger calls each forked an extractor subprocess,
fought for extract.duckdb's file lock, starved uvicorn, flipped the
container to unhealthy. Trigger now returns 409
(sync_already_in_progress) when held; _run_sync acquires non-blocking.
PARALLEL LEGACY FALLBACK
- Process pool fan-out for the _extract_via_legacy queue (default 8
workers, override via AGNES_KEBOOLA_PARALLELISM). Process pool, not
thread pool, because connectors/keboola/client.py:export_table does
os.chdir(temp_dir) — process-global, so threads raced and slice files
landed in the wrong directory ("[Errno 2] No such file or directory:
'<job_id>.csv_X_Y_Z.csv'").
- Extractor subprocess timeout 1800s -> 3600s (configurable via
AGNES_EXTRACTOR_TIMEOUT_SEC). 28+ tables × multi-minute Keboola export
jobs need the headroom on telemetry-class projects.
- Process group cleanup on timeout — Popen(start_new_session=True) puts
the extractor in its own group. On timeout the parent SIGTERMs the
group (10s grace) then SIGKILLs stragglers. Without this, the pool
workers were reparented to PID 1 and continued holding open Keboola
Storage export jobs. Inline extractor script also installs a SIGTERM
-> sys.exit(143) handler so the with ProcessPoolExecutor(...) block
__exit__ runs cleanly.
Tests: existing tests that patched subprocess.run updated to patch
subprocess.Popen with a _FakePopen stand-in (same exit-code-injection
contract). Two tests that exercised the parallel path forced
AGNES_KEBOOLA_PARALLELISM=1 to keep mocks alive (mocks don't ride into
ProcessPoolExecutor subprocesses).
Squashed onto current main (was 7 commits + multi-commit CHANGELOG +
agnes-auto-upgrade.sh conflicts; squash avoids per-commit conflict
resolution against main's flat-mount STATE_DIR refactor and 0.38.0
release cut).
* feat(keboola): Storage API direct extract path; drop extension data path
The DuckDB Keboola extension's COPY routes through Keboola QueryService,
which is unreliable on linked-bucket projects (extension v0.1.6 fixes
that case but isn't yet in the community CDN, and pre-fix any project
with the block-shared-snowflake-access feature flag couldn't see bucket
schemas at all). Move the extract path off the extension entirely and
talk to the Storage API directly via signed-URL download — works on any
project, regardless of extension state.
connectors/keboola/storage_api.py (NEW)
Lightweight client built on requests.Session. Three endpoints:
- POST /v2/storage/tables/{id}/export-async (kicks off job)
- GET /v2/storage/jobs/{id} (poll until done)
- GET /v2/storage/files/{id}?federationToken=1 (signed URL detail)
- GET <signed_url> (download bytes)
Supports sliced exports (manifest + per-slice signed URLs) and gzipped
payloads. ExportFilter dataclass mirrors the Keboola filter spec
(whereFilters / columns / changedSince / limit) and handles JSON
round-trip with the registry's source_query column. Token redaction
in error messages. Bounded exponential backoff on job polling.
No cloud-SDK dependency on the data path; thread-safe.
connectors/keboola/extractor.py
- materialize_query() rewritten: takes bucket/source_table/source_query
(JSON filter spec), exports via KeboolaStorageClient, converts CSV
to parquet via DuckDB, atomic os.replace. Same return shape so
sync.py downstream code stays uniform with the BQ branch.
- _extract_via_legacy() also moved to Storage API direct (kept the
name for caller compatibility with _legacy_worker / the parallel
batch extractor). Per-call temp directories — no os.chdir, threads
don't race.
app/api/sync.py
_run_materialized_pass for source_type='keboola' rows now constructs a
KeboolaStorageClient (replaces KeboolaAccess) and passes
bucket/source_table/source_query to materialize_query. Reuses one
client across rows for HTTP keep-alive. Sources keboola URL from env
too (KEBOOLA_STACK_URL) when instance.yaml doesn't have stack_url
configured.
cli/commands/admin.py
discover-and-register defaults Keboola rows to query_mode='materialized'
(NULL source_query = full table), matching the v26 migration's
unification of the local/materialized split for Keboola. BigQuery and
Jira keep their per-source defaults.
src/db.py
Schema bump 25 → 26. Migration: UPDATE table_registry SET
query_mode='materialized' WHERE source_type='keboola' AND
query_mode='local'. NULL source_query on those rows means "full table
export" — same effective behavior the local mode provided, but now
via Storage API instead of the extension.
pyproject.toml
kbcstorage dep stays (admin-side bucket/table list still uses the
SDK in app/api/admin.py / connectors/keboola/client.py); only the
data path is migrated off the SDK. Comment updated to reflect the
new boundary.
tests
- test_keboola_storage_api.py (NEW, 19 tests): ExportFilter parsing,
HTTP client (token redaction, retry logic, polling), download_file
(single, gzipped, sliced), end-to-end export_table_to_csv.
- test_keboola_materialize.py rewritten: mocks KeboolaStorageClient
instead of FakeAccess; same atomic-write + zero-rows + unsafe-id
contracts.
- test_sync_trigger_keboola_materialized.py: registry rows now carry
bucket+source_table+JSON-shape source_query.
114+ Keboola-impacted tests green locally.
* test: schema version assertion bumped to 26 alongside the keboola query_mode migration
* fix(keboola): cutover hot-patches surfaced on agnes-dev
Five small fixes that were applied as in-container hot-patches during
agnes-dev cutover and need to be on the source-of-truth image so a fresh
upgrade does not undo them.
- app/api/sync.py: auto-discover gate considers the WHOLE registry (any
source, any mode), not just rows where source matches and query_mode
is local. After the v25→v26 keboola materialized migration an
instance can have 30 materialized rows and zero local rows; the
previous gate kept re-firing _discover_and_register_tables every
scheduler tick, creating duplicate auto-discovered rows with the
wrong bucket prefix every time.
- app/api/admin.py: _discover_and_register_tables reassembles the
bucket as <stage>.<bucket-id> (e.g. in.c-finance) instead of
dropping the stage prefix; default query_mode for keboola is now
materialized (the v26 contract); validator allows NULL source_query
for keboola materialized rows (full-table export via Storage API
export-async, no SQL needed).
- cli/commands/admin.py: register-table mirrors the server validator
(NULL source_query allowed for source_type=keboola); --bucket help
text generalized to cover both BQ dataset and Keboola bucket id.
- connectors/keboola/extractor.py: max_line_size=64 MiB on
read_csv_auto so embedded JSON / SQL cells (kbc_component_configuration
in particular) do not trip the default 2 MiB ceiling.
- connectors/keboola/storage_api.py: GCP backend support — when the
Storage API returns a manifest whose slice URLs are gs://
references with a gcsCredentials block, rewrite to the JSON REST
download endpoint and authenticate with the issued OAuth bearer
token; redact tokens in any surfaced error string.
* test: align with new keboola materialized + auto-discover-gate contracts
- test_admin_keboola_materialized: rename
test_register_keboola_materialized_rejects_missing_source_query →
test_register_keboola_materialized_accepts_missing_source_query.
v25→v26 introduced 'keboola materialized with NULL source_query
means full-table export via Storage API export-async' as the
default registration shape; the rejection case is no longer the
contract.
- test_sync_filter: add list_all() to _StubRegistry. The auto-discover
gate in _run_sync now keys off the WHOLE registry (not just local
rows) so materialized-only Keboola instances do not re-trigger
discovery on every tick.
* feat(keboola): native parquet export — skip CSV roundtrip
Storage API export-async accepts fileType={csv,parquet}. Switching the
materialized sync to parquet eliminates the CSV → DuckDB COPY → parquet
roundtrip that pinned a single uvicorn worker over 4 GiB on multi-GB
tables (read_csv with all_varchar + max_line_size=64MB has to
materialize the whole CSV in memory before COPY can stream out a
parquet). Snowflake UNLOAD on Keboola's side already produces typed,
self-contained parquet files; the extractor downloads them and renames
into place.
Two cases:
- **Single-file** export (small table): file_info.url points at one
signed URL; download_file streams chunks straight to .parquet.tmp
and we're done. No DuckDB.
- **Sliced** export (Snowflake UNLOAD respects MAX_FILE_SIZE — 16 MiB
default — so anything larger arrives as N parquet slices): each
slice is a complete parquet file with its own footer; naive concat
would corrupt them. download_file_slices keeps the slices as
separate files in a tempdir, then DuckDB COPY (SELECT * FROM
read_parquet([slice0, slice1, ...])) merges them into one
consolidated parquet. DuckDB streams row groups during this — peak
memory bounded to one row group (~1 MiB) regardless of source size.
The legacy CSV path stays as the explicit opt-in via source_query=
'{"file_type":"csv"}' for projects whose backend can't UNLOAD
parquet (none known today; cheap escape hatch). Backward-compat alias
KeboolaStorageClient.export_table_to_csv kept.
Also fixes a latent bug in download_file's gzip detection: previous
heuristic flagged any unencrypted file as gzipped, which would have
corrupted parquet downloads at gunzip time. Name-suffix-only now.
* fix: tempdir leak cleanup, every 0m schedule, /sync/trigger body shapes
Three small self-contained fixes uncovered during agnes-dev cutover.
- connectors/keboola/extractor.py: tempfile.TemporaryDirectory now uses
ignore_cleanup_errors=True so a worker death mid-write doesn't leave
multi-GiB stale slice trees on the boot disk. (12 GiB seen after a
disk-full crash where TemporaryDirectory's own cleanup also raised
and got swallowed.)
- src/scheduler.py: is_valid_schedule accepts 'every 0m' (interval=0
= always due). Force-resync of an errored row no longer requires
waiting out the default 'every 1h' interval — admin can flip the
schedule, trigger, then flip back.
- app/api/sync.py: POST /api/sync/trigger accepts both ['table_id']
(legacy bare-array body) and {'tables': ['table_id']} (matches the
response payload shape, more discoverable for clients building
requests by hand). Malformed bodies return 422 with a structured
detail; null/missing means 'sync everything' as before.
Tests cover: tempdir cleanup on raise (sliced parquet path),
is_valid_schedule + is_table_due 'every 0m' acceptance, and trigger
body parametrized matrix (8 valid shapes + 6 rejection cases).
* fix: targeted-trigger filter in materialized pass + auto-upgrade defer
Two operational gaps observed during agnes-dev cutover, in the same
sync-routing area.
- _run_materialized_pass now takes a 'tables' arg and skips rows not in
the target set with reason='not_in_target'. POST /api/sync/trigger
with a body of tables previously only scoped the legacy extractor
subprocess — the materialized pass kept iterating every due
materialized row, so an admin asking to re-sync kbc_job re-ran
every other due materialized row alongside it. Match on registry id
OR name (admins commonly pass either form). tables=None preserves
the no-filter behavior.
- New GET /api/sync/status (public, no auth) returns {locked: bool}
off _sync_lock.locked(). agnes-auto-upgrade.sh probes this before
docker compose up -d and exits 0 with a 'deferred recreate' log
line if a sync is in flight — the next 5-min cron tick retries.
Pre-fix, an auto-upgrade triggered mid-sync would recreate the
uvicorn worker and kill the in-flight extractor / Snowflake-UNLOAD
download (observed when kbc_job's first 7-day retry got SIGKILLed).
Connection failures in the probe fall through to the upgrade —
being stuck on a wedged image is worse than interrupting a
hypothetical sync.
* fix: auto-discover protects admin overrides + surfaces drift
Two real-world incidents on agnes-dev drove this:
1. kbc_job was registered manually with the correct
(in.c-kbc_telemetry, kbc_job) coordinates. A naive auto-discover
re-run would have inserted a SECOND kbc_job row at the slugified
id 'in_c-keboola-storage_kbc_job' (where Keboola's discovery
places it) — and that row's Storage API export-async 404s.
2. An earlier auto-discover bug stripped the stage prefix from
bucket ids ('c-finance' instead of 'in.c-finance'), inserting
137 rows whose syncs all failed.
Fix:
- _discover_and_register_tables now builds a plan first
(_build_keboola_discovery_plan) classifying each discovered table
into one of new / existing_match / existing_drift / invalid, then
executes only the 'new' bucket. Drift rows are reported with both
sides of the disagreement plus drift_kind:
- same_id_diff_coords: registry has the same id but different
bucket / source_table (admin migrated coords inline).
- name_collision: discovery's slugified id differs from any
registry id, but the discovered .name matches an existing row's
.name (case-insensitive). Catches the kbc_job case.
- Bucket detection now prefers the API's authoritative bucket_id
field (separate field on the Keboola tables.list response,
normalised by KeboolaClient.discover_all_tables). Falls back to
id-string parsing only when bucket_id is missing (older fallback
path inside discover_all_tables).
- Endpoint POST /api/admin/discover-and-register?dry_run=true
returns the plan without writing — would_register, drift,
invalid lists. Lets an operator audit before merging discovery
with a registry that has admin overrides.
Removed 'every 0m' from test_register_request_rejects_malformed_sync_schedule
— the runtime started accepting it in the previous commit (force-resync
override) and the validator follows suit.
* feat(keboola): AGNES_TEMP_DIR routes tempfiles off overlayfs /tmp
The container's /tmp lives on the boot disk's overlayfs (29 GiB on
agnes-dev, shared with /var). Snowflake UNLOAD of a wide table writes
slices into per-call /tmp tempdirs that fill multi-GiB / many-slice
exports long before the dedicated data disk fills. agnes-dev hit
100% boot-disk while the 20 GiB data disk had 15 GiB free.
connectors.keboola.storage_api.get_temp_root() reads AGNES_TEMP_DIR;
mkdirs the target on first use; unset / empty / unwritable falls
back to None (system tempdir, OSS-pre-fix behaviour). Both
materialize_query (parquet path) and _extract_via_legacy (CSV
fallback) and the sliced-CSV concat path in storage_api use the
helper now.
docker-compose.yml defaults AGNES_TEMP_DIR=/data/tmp on app, scheduler,
and extract services. The data volume is the dedicated disk in
production layouts and a plain docker volume in single-disk
dev/laptop setups — same blast radius as the previous /tmp default
on the latter, no regression.
Operator-and-analyst quality bundle: a security fix for the optional
Telegram bot, two CLI gaps closed, and three rounds of UX polish on
`agnes diagnose` and `agnes pull` so non-TTY consumers (CI runners,
Claude Code SessionStart hooks, sub-agent watchdogs) get readable,
actionable signal.
- Pairing-code RNG: random.choices -> secrets.choice (CSPRNG).
- Telegram script runner: refuse out-of-shape usernames before sudo -u.
CLAUDE.md.bak.<ISO-timestamp> before regenerating.
- agnes admin unregister-table <id> -> DELETE /api/admin/registry/{id}
- agnes admin update-table <id> --field=value ... -> PUT /api/admin/registry/{id}
response but never promotes the headline. BQ billing-equals-data check
downgraded warning -> info.
default (5 s / 1 MiB vs 30 s / 10%) so sub-agent watchdogs don't kill
the pull as a hung process. New env knobs:
AGNES_PULL_PROGRESS_INTERVAL_{SECONDS,BYTES}.
--include-schema (or ?include=schema) to opt back in.
Tests: 120 passed across the touched modules, including new tests for
each fix. Pre-existing failures on main (DB migration v1->v9, binary
rename) are unrelated and not introduced here.
Hook JSON path uses /reload-plugins (no restart needed); manual-mode
echo path was still telling the operator to /exit + restart. Both now
say /reload-plugins.
Tests renamed to *_reload_hint_* to match the new wording.
Claude Code's `/reload-plugins` slash command picks up newly installed
plugins into the running session without forcing the user to /exit and
restart Claude Code. The hook JSON `systemMessage` and `additionalContext`
both now point at it.
Tests updated to pin the new hint shape.
Adds end-to-end flow for installing and keeping the per-user filtered
Claude Code marketplace in sync with the user's Agnes stack
(admin RBAC grants \ MyAIStack opt-outs U /store installs).
Setup (one-liner in install prompt step 5):
`agnes refresh-marketplace --bootstrap` clones the per-user marketplace
bare repo to ~/.agnes/marketplace, strips PAT from the cloned origin
URL, registers the local path with Claude Code, and installs every
plugin in the served manifest at --scope project. Replaces a 15-line
inline shell sequence that tripped Claude Code's agent-driven `rm -rf`
permission gate.
Auto-refresh (SessionStart hook installed by `agnes init`):
`agnes refresh-marketplace --quiet` runs every Claude Code session,
fetches+resets the clone (server rebuilds as orphan commits, so
pull --ff-only is impossible), and version-aware reconciles:
- missing in workspace -> claude plugin install <name>@agnes --scope project
- version differs -> claude plugin update <name>@agnes
- matches -> skip
Don't auto-uninstall plugins that disappeared from the manifest --
a transient empty manifest from the server would wipe the stack.
Hook output: when --quiet AND something actually changed, emits Claude
Code hook JSON on stdout -- `systemMessage` (transient toast) and
`hookSpecificOutput.additionalContext` (model-side system reminder),
both carrying the change summary plus a "/exit + restart Claude Code"
instruction (Claude only scans plugins at session start).
Windows hook compatibility: the refresh-marketplace hook command is
wrapped in `bash -c "..."` because Claude Code on Windows runs hook
commands directly without invoking a shell, so `2>/dev/null || true`
would otherwise be passed as literal argv tokens.
Cross-cutting:
- cli/lib/marketplace.py: shared CLONE_DIR + MARKETPLACE_NAME constants.
- cli/lib/hooks.py: SessionStart now has two independent entries
(pull + refresh-marketplace) so a failure in one doesn't suppress
the other; legacy `da sync` and prior single-pull layouts upgrade
cleanly on re-init.
- PAT injection on every git fetch via per-invocation credential
helper (token in \$AGNES_TOKEN env, never in argv or .git/config).
- Pre-snapshot of installed plugins captured BEFORE
`claude plugin marketplace update` so silent auto-applied version
bumps still fire notifications.
- scripts/dev/agnes-client-reset.sh: cleans ~/.claude/plugins/marketplaces/agnes,
~/.claude/plugins/cache/agnes, drops uv build cache, documents
workspace-scoped residue that can't be enumerated from the script.
- app/web/setup_instructions.py: legacy AGNES_DEBUG_AUTH path also
uses clone (direct HTTPS marketplace add is broken end-to-end on
every Claude Code distribution -- stores response as single file,
plugin source paths then 404).
28 new tests (test_cli_refresh_marketplace.py) + extended hook + setup
template tests cover bootstrap, fetch+reset ordering, version-aware
reconcile, project-path filtering, hook JSON shape, and the bash-c
Windows wrapper invariant.
Single hook entry: 'agnes self-upgrade --quiet ... || true; agnes pull
--quiet ... || true'. Shell semicolon guarantees ordering across every
Claude Code version (no reliance on undocumented multi-hook execution
semantics); each segment's || true preserves the original property
that an upgrade failure does not abort the pull.
Reuses cli.update_check.check() for the version probe — extended with
bypass_disabled=True so explicit user-typed self-upgrade is not silenced
by AGNES_NO_UPDATE_CHECK (which is for the implicit warning loop).
Install path: uv tool install --force when uv is on PATH; otherwise
curl + pip via sys.executable (NOT system python3, NOT --user — both
would land outside the agnes venv and silently no-op the upgrade).
Smoke test execs the binary at the install-resolved path (uv tool dir
joined with agnes-the-ai-analyst/bin/agnes, or sys.executable's sibling
agnes for pip) — never via shutil.which, which can resolve a stale shadow
on PATH and produce a false-positive smoke pass on the OLD version. Smoke
also asserts --version output contains info.latest via PEP 440 Version()
equality (so 0.40.0 does not falsely match 0.40.10).
On smoke fail: rollback to last_known_good.json (written only after a
previous run's smoke passed). Rollback rc is captured and surfaced on
stderr if it also fails. First-ever upgrade or unrecoverable rollback
prints the canonical bootstrap recovery: curl -fsSL <server>/cli/install.sh | bash.
AGNES_SELF_UPGRADE_IN_PROGRESS=1 is set for the duration of the run
and propagated to the smoke-test subprocess. Layer B's _check_version_headers
honors the sentinel and skips the < min hard-stop, so an in-flight
upgrade can never sys.exit(2) itself.
--force invalidates the update_check cache BEFORE probing. --force +
offline = exit 1 with explicit stderr (without --force, offline is silent).
--quiet suppresses progress output but never gags failure stderr.
Every API response is inspected via httpx event_hooks. When the server
reports X-Agnes-Min-Version > local, CLI prints a remediation message
and exits 2. Latest-version drift continues to be handled by the
update_check warning loop — no double-warning on every API call.
Adds X-Agnes-Latest-Version and X-Agnes-Min-Version headers to every
/api/* response. CLI consumes these to hard-stop on incompatible drift.
MIN_COMPAT_CLI_VERSION ships at 0.0.0 — no enforcement until a deliberate
wire-protocol break bumps it.
Also dedupes app version logic: app/main.py:_app_version() helper deleted,
replaced by app/version.py:APP_VERSION as the single source of truth.
test_app_version.py rewritten to target app.version.
`_rewrite_user_sql_for_bigquery_query` does its own bare-name detection
(mirroring the non-RBAC parts of `_bq_guardrail_inputs`). The backtick
masking from #201 was applied to `_bq_guardrail_inputs` and the
forbidden-table loop, but missed this third site — so a registered
local-mode table name appearing as the table segment of a
user-supplied full backtick path (e.g. ``\`prj.ds.orders\`` matching
registered local ``orders``) tripped the cross-source guard and
forced every backtick-path query into the 50-100× slower
ATTACH-catalog fallback.
Mask once at the top of the function, route both the BQ-name
detection (line ~830) and the cross-source check (line ~867) through
the masked copy. New regression test
`test_local_name_inside_backtick_path_does_not_trip_cross_source`
proves the wrapper now wraps when it should.
Adds Pass 3 to `_bq_guardrail_inputs` that scans user SQL for full
backtick paths `<project>.<dataset>.<table>` and gates them
identically to the `bq."<dataset>"."<table>"` pass:
- Project must match the configured BigQuery data project
(`get_bq_access().projects.data`). Mismatch → HTTP 403
`bq_path_cross_project`.
- Path must point at a registered row. Unregistered → HTTP 403
`bq_path_not_registered`.
- Non-admin caller must hold a grant on the registered row's id.
Missing grant → HTTP 403 `bq_path_access_denied`.
Pre-fix, full backtick paths bypassed Agnes RBAC entirely — only the
service account scope limited reach. Post-fix the boundary matches
what `agnes catalog`-driven flows already enforce. Admin still
bypasses the per-id grant check but cannot bypass registration or
project match.
Pass 3 also seeds `dry_run_set` for resolved registered paths so the
cost-cap dry-run runs against the same physical table the user named
— composing cleanly with the Layer 2 fail-fast fallback.
When BQ rejects the rewritten dry-run SQL with `bq_bad_request`, the
cap-guard now retries with the user's ORIGINAL SQL instead of building
a synthetic `SELECT * FROM <table>` per registered table. The
synthetic path threw away user filters / projections / partition
predicates and routinely ballooned the estimate to "full table size",
falsely tripping `remote_scan_too_large` on legitimate narrow queries
(typical issue #201 trace: rewriter corrupts a backtick path → BQ
parse error → synthetic over-estimate → 400).
Behaviour:
- Rewritten SQL succeeds: same as before (issue #171 single-dry-run).
- Rewritten SQL parse-errors, original SQL succeeds: use original
estimate. Common case for users submitting BQ-native input.
- Both fail with `bq_bad_request`: HTTP 400 `remote_estimate_failed`
with a hint pointing at `agnes catalog` / BQ-native syntax. No
silent over-estimate.
- Non-parse BQ error (forbidden, upstream): still 502 as before.
This is a behaviour change for clients matching error kinds — failure
to estimate scan size now surfaces as `remote_estimate_failed`
instead of being masked behind `remote_scan_too_large` from the
synthetic path.
Replaces the existing `test_guardrail_falls_back_to_per_table_estimate_on_bq_parse_error`
(which pinned the old contract) with `test_fallback_tries_original_sql_first`
and `test_fallback_fails_fast_on_pure_duckdb_syntax`.
`agnes query --remote` corrupted user SQL when the request contained a
full BigQuery backtick path (`<project>.<dataset>.<table>`) whose
table segment matched a registered bare-name alias. The bare-name
rewriter used `\b` word-boundary matching against the lower-cased SQL;
both `.` and `` ` `` are non-word characters, so the regex fired
INSIDE the user's backtick path and produced malformed nested-backtick
SQL that BigQuery rejected at parse time.
Fix:
- Add `_mask_backticks(sql)` helper: replace each `…` segment with
spaces of equal length, preserving offsets so word-boundary
searches find positions only outside backticks.
- `_bq_guardrail_inputs` (bare-name pass + forbidden-table pass)
searches against the masked SQL.
- `_rewrite_bq_table_refs_to_native` Pass 1 splits the SQL on
`(\`[^\`]*\`)` and rewrites only the outside-backtick chunks. Pass
2 (`bq."ds"."tbl"` → backtick form) is unchanged — its prefix can't
appear inside backticks.
Adds three regressions covering the rewrite + guardrail paths.
Pre-existing test_orchestrator_skips_orphan_parquet_in_extracts caught
the regression: my filesystem fallback created master views for ANY
parquet on disk, including orphans where DELETE /api/admin/registry
removed the registry row but the parquet wasn't fully cleaned up.
Fix: load the set of registered materialized table_ids for THIS source
from table_registry before the scan, and skip any parquet whose stem
isn't in that set. If the registry read fails (test fixture, transient
DB error), skip the fallback entirely — orphan exposure is worse than
missing master view recovery.
Pre-existing test now passes. New regression test pins the orphan-skip
contract specifically for the filesystem-fallback path.
0.40.0 added _persist_materialized_inner_view in materialize_query, which
tried to open extract.duckdb from a fresh DuckDB handle to write the _meta
row + inner view. In production this conflicts with the same uvicorn
process's existing read-only ATTACH (orchestrator's analytics conn holds
extract.duckdb ATTACHed as <source_name> alias), and DuckDB single-process
file-handle uniqueness rejects with:
Binder Error: Unique file handle conflict: Cannot attach "extract"
— already attached by database "<source>"
The helper logs WARNING fail-soft, parquet stays canonical, but the
master view never appears via the meta path.
Fix: at the end of _attach_and_create_views, scan
<extract_dir>/data/*.parquet and CREATE OR REPLACE VIEW <id> AS
SELECT * FROM read_parquet('<path>') for any parquet whose <id> is not
already in the per-source tables list (= meta path didn't pick it up).
Decoupled from materialize_query open-handle race. Honors the same
view_ownership cross-connector collision rules as the meta path
(first-come-first-served via view_repo.claim).
Tests:
- filesystem-fallback fires when _meta row missing
- skipped when meta path already created the view (no shadow)
- skips invalid identifiers (e.g. parquet stem starting with a digit)
- doesn't crash when source has no data/ subdir
Pre-fix flow:
1. extractor subprocess writes _meta with N remote rows + creates N inner
views in extract.duckdb (rebuild_from_registry skips materialized rows
per design — explicit `continue` at line 389)
2. _run_materialized_pass calls materialize_query, which writes parquet
atomically + returns stats — but never updates _meta
3. orchestrator.rebuild scans _meta, finds only the N remote rows, creates
master views only for them. Materialized parquet is on disk but
invisible to /api/query → 400 'not yet materialized'
Symptom appears after every container recreate (the previous run's _meta
state is wiped because docker compose down nukes the named volume that
backs extract.duckdb on some compose layouts; even on volumes that
persist, the next extractor pass calls _create_meta_table which DROPs
+ CREATEs _meta cleanly).
Fix: after os.replace(tmp_path, parquet_path) in materialize_query, open
extract.duckdb (read-write), DELETE existing _meta row for table_id,
INSERT new one with query_mode='materialized', and CREATE OR REPLACE
VIEW <table_id> AS SELECT * FROM read_parquet(<path>). All inside a
single transaction so concurrent reads see either old or new state, not
torn rows. Fail-soft on lock contention or schema drift — parquet
remains canonical, next sync pass recovers.
Tests: 3 new in test_bq_materialize.py covering:
- meta + inner view registered after materialize, alongside existing
remote rows
- re-run replaces (not duplicates) the meta row
- skips inner-view registration when extract.duckdb doesn't exist yet
(fresh BQ-only deployment edge case)
In cross-project BQ setups (where billing != data), the SA typically has
serviceusage.services.use on the billing project but not on the data
project. The rewriter passed bq.projects.data as the first arg to
bigquery_query(), which BQ uses as the execution + billing project →
403 USER_PROJECT_DENIED.
Match the convention used everywhere else in the codebase
(app/api/v2_scan.py, app/api/v2_sample.py, app/api/v2_schema.py,
connectors/bigquery/extractor.py): backtick paths inside the inner SQL
use the **data** project (resolves the actual table location), the
bigquery_query() first arg uses the **billing** project (decides who
pays + which project the job runs under). For single-project deploys
the two are identical so the fix is a no-op there.
Test pins the cross-project case: data-prj for backticks, billing-prj
for the bigquery_query() first arg.
R3 final pass surfaced one issue, addressed:
R2#2 introduced PID-suffixed <target>.{pid}.tmp / .{pid}.partN to
prevent concurrent agnes pull invocations from yanking each other's
in-progress writes. The pre-clean inside _download_chunked /
_download_single_stream only deletes leftovers from the CURRENT
process's PID — files from a SIGKILL'd or crashed prior pull (any
other PID) are never touched and accumulate on disk forever.
Add _reap_dead_pid_leftovers(target_path) called at the start of both
download paths. Globs <target>.*.tmp / <target>.*.partN, extracts the
embedded PID, calls os.kill(pid, 0) to test liveness (POSIX standard
no-op probe), and unlinks files whose process no longer exists.
Permission-denied = process is alive but owned by another user → keep
the file (conservative). Windows users get the conservative 'keep'
default.
Two new tests pin the behavior — live-PID file preserved, dead-PID
.tmp + .partN reaped, bare-name (legacy) untouched, garbage filenames
skipped without raise.
R1 adversarial review surfaced 5 issues, all addressed:
#1 chunked download silently disabled in non-Caddy deployments (HEAD on
GET-only FastAPI route returns 405). _probe_range_support now falls back
to GET with Range: bytes=0-0 when HEAD fails — works against both
Caddy file_server (HEAD-friendly) and dev FastAPI direct (GET-only).
#2 parse-error fallback heuristic too broad — matched on Unrecognized
name / Function not found / No matching signature / Invalid cast,
which BQ surfaces for ordinary user-column typos. That triggered slow
ATTACH-catalog retry on every typo (2× latency tax). Narrowed to just
'Syntax error' / 'syntax error' which are the genuine DuckDB-vs-BQ
dialect mismatch markers.
#3 apply_bq_session_settings was only run on fresh-built pool entries,
not on reuse. An operator's /admin/server-config change to bq_query
_timeout_ms wouldn't propagate to long-lived pooled sessions until
restart. Fixed: re-apply on every pool acquire (idempotent + fail-soft).
#4 content-length sanity bound — a misconfigured proxy returning a
wildly inflated Content-Length would cause overlapping chunked Range
requests against the actual file → corrupt assembled output (caught
by manifest hash check, but only after wasted bandwidth). Cap at 100
GiB; above that, drop to single-stream.
#5 rewriter assumed every BQ row resolves under the single
bq.projects.data project. Bucket containing '.' suggests a project-
qualified bucket (multi-project deployment); rewriter would silently
target the wrong project. Conservative skip with regression test.
Address code-reviewer findings on the bigquery_query() rewrite path:
1. Outer LIMIT wrap — bigquery_query() materialises BQ result into DuckDB
before fetchmany sees it (vs ATTACH-catalog Storage Read API streaming).
A user 'SELECT *' against a billion-row remote table would buffer the
entire result before request.limit applied. Wrap rewritten SQL in an
outer 'LIMIT N+1' so the cap pushes into the BQ job itself.
2. Dollar-quoted inner SQL — naive replace("'", "''") doubling missed
DuckDB backslash-escape sequences (\\, \\n, \\t, …). A predicate
like 'WHERE name = ''O\\'Brien''' was unsafe under the doubling
path. DuckDB $bqq_inner$ … $bqq_inner$ form takes the inner SQL
verbatim with no escapes whatsoever. Falls back to legacy doubling
if user SQL improbably contains the literal tag.
3. Parse-error fallback — when the rewritten path fails with a BQ-side
parse / validation error (DuckDB-only syntax like ::INT cast that
survives identifier rewrite but BQ refuses), retry the user's
original SQL via the legacy ATTACH-catalog path so the request still
succeeds. Mirrors the existing dry-run fallback contract.
4. CHANGELOG — delete duplicate CLI bullets that landed under
already-released [0.38.1] (file corruption from merge — entries are
correctly under [0.39.0]).
Two improvements to `agnes pull` progress reporting:
1. **Aggregated per-file progress across chunked downloads**: the
existing Rich progress bar already used one task per file, but the
chunked-download contract (one file = N parallel chunk callbacks
summing to file size) meant we needed to verify that all chunk
threads advance the same task. They do — the per-file callback is
constructed once per tid and routes every chunk's byte delta to the
same task / textual entry, so the bar shows one aggregated bytes-
downloaded total rather than N separate sub-bars.
2. **Textual fallback for non-TTY stderr**: when stderr is not a
terminal (SessionStart hook, CI runner, Docker log capture), Rich
either suppresses output (silent multi-minute pull on a 5 GB
parquet) or emits raw control sequences. The new `_TextualProgress`
helper instead emits one plain-text line per file at most every
10%-of-total-bytes or 30 s, plus a final `100% done` line per file.
Format: `[N/T files] <tid>: 25% (16 MB / 66 MB) at 1.5 MB/s`.
The TTY path is unchanged. Detection uses `sys.stderr.isatty()` —
`show_progress=True` flips into the textual fallback when that returns
False. `show_progress=False` (the SessionStart hook) still emits no
progress text in either mode.
translate_bq_error previously mapped BQ's responseTooLarge failure mode
to bq_bad_request (HTTP 400 with the raw upstream message). The user-
facing implication ('your SQL has a syntax error') is wrong -- the root
cause is query shape (BQ refused to return the result inline because
it exceeded the response size limit), and the actionable remediation is
'narrow the WHERE clause, aggregate further, or use a materialized
table'.
Add bq_response_too_large as a first-class BqAccessError kind (also 400)
with a canonical hint message; original BQ message preserved in details
for operator debugging. Detection is substring-based on 'response too
large' and fires before the generic BadRequest path so the dedicated
mapping always wins. Affects every BQ-touching path since they all
share translate_bq_error -- /api/query, /api/v2/{scan,sample,schema},
materialize.
Pool the httpx.Client used by `stream_download` so N parquet downloads
share a single TLS handshake instead of one handshake each. With the
optional `h2` package installed, HTTP/2 multiplexing further lets all
chunk Range requests share a single TCP connection — synergizes with
the range-chunked download path added in the previous commit.
The shared client is created lazily on first stream-download call, kept
alive for the duration of the process via a module-level slot, and
closed at exit via `atexit.register`. Construction wraps in a
try/except: when `h2` is unavailable (slim install), httpx raises
ImportError on `http2=True` and we transparently fall back to an
HTTP/1.1 client — pooling alone still amortizes TLS handshakes.
`agnes pull` must never crash on a missing optional package, so the
fallback path is non-negotiable. `h2>=4.1.0` is added to the core
dependency set; downstream slim installs that drop it lose the HTTP/2
benefit but keep correctness.
Each BqAccess.duckdb_session() acquire previously created a fresh
in-memory DuckDB conn and ran INSTALL bigquery; LOAD bigquery;
CREATE SECRET; ATTACH on it -- costing ~0.5 s per request even before
any BQ work. Add a process-local pool (deque + lock) of pre-warmed
sessions; acquire reuses a warm entry when available, refreshing the
auth SECRET so a long-lived pool entry doesn't keep a stale GCE
metadata token past its TTL. Liveness probe (cheap SELECT 1) drops
broken entries before handing them to callers.
On exception inside the with-block the conn is closed instead of
returned to pool (session may carry dirty state). Pool size is
data_source.bigquery.session_pool_size (default 4; sentinel 0
disables pooling). Process-cached, not fork-safe (single uvicorn
worker is the supported deployment shape per CLAUDE.md).
All call sites get faster automatically: /api/query, /api/v2/{scan,
sample,schema}, materialize, the orchestrator's remote-attach, and
the BQ dry-run cap-guard.
When the server advertises `accept-ranges: bytes` and a parquet exceeds
`AGNES_PULL_CHUNK_THRESHOLD_BYTES` (default 50 MB), `stream_download`
now splits the file into N parallel HTTP Range requests
(`AGNES_PULL_CHUNK_PARALLELISM`, default 4, capped 1..16) and
assembles the parts into the destination atomically.
Targets the per-flow-shaped network (corp VPN with per-TCP-connection
rate-limiting) where single-stream throughput is throttled but N parallel
streams over the same connection scale roughly linearly. Manifests with
1 large materialized parquet + N remote tables previously left the
existing across-files `AGNES_PULL_PARALLELISM=4` pool with 1 active
worker = single-stream throughput; this fixes that.
Falls back to single-stream when:
- HEAD doesn't advertise `accept-ranges: bytes`
- Server returns 200 instead of 206 to a Range probe
- File size below the threshold
Cleanup discipline: every part file removed before return (success or
failure); destination written via `<target>.tmp` and renamed atomically.
Per-chunk retry on transient network blips (bounded by AGNES_STREAM_RETRIES).
User SQL hitting query_mode='remote' BigQuery rows was 50-100x slower
than the equivalent direct bigquery_query() call because DuckDB's master
view (CREATE VIEW … AS SELECT * FROM bigquery.<ds>.<tbl>) does not push
WHERE/SELECT/LIMIT into BQ in ATTACH-catalog mode. The BQ extension opens
a Storage Read API session over the entire upstream table; on >100M-row
sources this was 70-150s and frequently failed with 'Response too large
to return'.
Extract the existing dry-run rewriter's core (table-name → BQ-native
backtick path) into a shared helper. Add an execution-path rewriter
that wraps the whole user SQL in bigquery_query('<project>', '<inner>')
so the BQ planner sees the full query and engages partition pruning +
projection pushdown server-side.
Conservative fall-through: cross-source JOINs (BQ ↔ Keboola/Jira local),
queries already containing bigquery_query(, and unconfigured BQ project
all skip the rewrite and run the original SQL via ATTACH-catalog so
behavior degrades gracefully.
The DuckDB BigQuery extension defaults bq_query_timeout_ms to 90 s,
which is too tight for analyst-scale queries against view-backed BQ
datasets. Agnes already has apply_bq_session_settings() that bumps it
to 600 s (configurable via data_source.bigquery.query_timeout_ms), but
two regressions let the 90 s default leak through to live queries:
1. apply_bq_session_settings() swallowed every Exception silently. If
the BigQuery extension wasn't loaded on the connection yet, or the
installed extension version didn't recognise the setting, the SET
would fail and the function would return without surfacing the
problem. Operators saw 90 s timeouts on 'agnes query --remote' with
no log line explaining why.
2. The call sites in src/db.py:_reattach_remote_extensions and
src/orchestrator.py:_remote_attach only invoked
apply_bq_session_settings on the metadata-token branch (token_env
empty, the BqAccess contract). The token-based and no-auth branches
ran ATTACH against the BigQuery extension without ever applying the
timeout setting — so any BQ source registered with an explicit
token_env, or with no auth env at all, fell back to the 90 s default.
Fix:
- apply_bq_session_settings now logs WARNING on each failure path
(instance_config import error, non-numeric value, SET execution
failure, readback error). It also verifies the setting actually
landed via SELECT current_setting('bq_query_timeout_ms') and logs
WARNING when the readback disagrees with the requested value, which
catches the silent-ignore case some extension versions exhibit.
- Both _reattach_remote_extensions (src/db.py) and _remote_attach
(src/orchestrator.py) now call apply_bq_session_settings on every
branch that ATTACHes a BigQuery alias, not only the metadata-token
branch. Idempotent: calling it twice on the metadata-token path is a
no-op SET.
Tests:
- Extended the _RecordingConn fixture to support .fetchone() so the
readback assertion path works. Updated existing call-shape
assertions to expect the SELECT current_setting readback alongside
the SET. Added two new tests covering the WARNING surfaces for SET
failure and readback mismatch — regression guards for the silent-
fallback bug this PR addresses.
- Full BQ-touching suite (398 tests) passes.
Five hottest BQ-touching endpoints were `async def` but invoked synchronous
DuckDB / BQ-extension calls inside the body. Under uvicorn's single event
loop that meant a single heavy `agnes query --remote` (waiting up to
~200 s for BQ's jobs.query) froze EVERY other request — /api/health,
dashboard, auth, even another query — for the full BQ wait. Operators
saw "VM idle, app frozen" during PR #188's testing.
Convert to plain `def` so FastAPI auto-offloads the body to the anyio
thread pool. Event loop stays free for non-BQ requests.
- app/api/query.py:execute_query
- app/api/v2_scan.py:scan_estimate_endpoint, scan_endpoint
- app/api/v2_sample.py:sample
- app/api/v2_schema.py:schema
Audit: 0 `await` statements in any converted handler (verified file-by-
file), so the rename is safe. Tests in tests/test_v2_*.py called the
handlers via `asyncio.run(...)` which now fails on a non-coroutine return;
swapped for direct calls (asyncio.run( -> ( ) — keeps paren balance).
Plus AGNES_THREADPOOL_SIZE env var (default 200, was anyio's stock 40)
in app/main.py:lifespan. Set via
anyio.to_thread.current_default_thread_limiter().total_tokens. 200 is
comfortable headroom for <50 concurrent analysts; bump for more.
480/480 impacted tests pass (the 2 remaining errors are a pre-existing
fixture setup issue in test_reader_smoke_matrix.py unrelated to this
change).
A single analyst's multi-GB `agnes pull` held the only uvicorn worker
for the duration of the stream, starving UI / /api/health / every other
API endpoint. Container flipped to `unhealthy`. Triggered while a
6.8 GB `order_economics` pull was in-flight on prod 2026-05-05.
Caddy now intercepts `GET /api/data/{table_id}/download` and serves
the parquet directly via sendfile from the data volume (mounted r-o
at /srv inside the caddy container). RBAC enforced by `forward_auth`
to a new lightweight `GET /api/data/{table_id}/check-access` endpoint
(returns 204 / 403) — the bulk transfer never reaches uvicorn.
Path discovery via `try_files` over the known extract.duckdb v2 source
subdirs. Anything not at a static path falls through to the existing
app handler so legacy `src_data/parquet` and future connectors still
work without a Caddyfile change. Non-Caddy deployments are unchanged.
Stage 1 (multi-worker uvicorn) was considered but blocked by the
single-writer DuckDB lock on system.duckdb — workers > 1 would crash
at startup on "Could not set lock on file", the same race that pushed
the scheduler from in-process writes to HTTP-via-app. Multi-reader
workers + single-writer coordination is out of scope for this PR.
DuckDB BigQuery extension defaults `bq_query_timeout_ms` to 90 s, which
is too tight for analyst-scale queries against view-backed BQ datasets.
`agnes query --remote` HTTP 400'd with `Binder Error: Query execution
exceeded the timeout. Job ID: ...` whenever the underlying BQ job ran
longer than 90 s, even though the job itself was healthy.
Add `data_source.bigquery.query_timeout_ms` (default 600 000 ms = 10 min,
sentinel 0 falls through to the extension default). Applied via
`SET bq_query_timeout_ms` after every `LOAD bigquery` on every BQ-touching
DuckDB session: orchestrator's `_remote_attach` ATTACH path, BqAccess
session factory, and the standalone extractor. Configurable via
`/admin/server-config` UI.
Fail-soft: extension versions that don't recognise the setting silently
keep the default rather than poisoning the session.
* fix(keboola): per-table fallback to legacy Storage-API client
The DuckDB Keboola extension's per-table COPY fails with
`Schema '..."in.c-..."' does not exist or not authorized` on
projects whose Snowflake backend doesn't expose bucket schemas
to the storage-token-derived QueryService role
(keboola/duckdb-extension#17). ATTACH itself succeeds, so the
existing extension-level fallback in `_try_attach_extension`
never triggers — the table is just marked failed.
- Promote `kbcstorage>=0.9.0` from optional to core dep so the
legacy client import in `_extract_via_legacy` doesn't crash
default installs with `ModuleNotFoundError`.
- Wrap `_extract_via_extension` in a per-table try/except so a
scan failure retries via `_extract_via_legacy` instead of
recording `tables_failed` and moving on.
Slower than the extension path, but produces correct parquets
on affected projects while the upstream extension fix lands.
* test(keboola): cover per-table extension→legacy fallback
Two existing tests mocked _extract_via_extension to throw and asserted
the original message survived in result["errors"]. With per-table
fallback, the new flow retries via _extract_via_legacy — which on the
mock URLs would throw a different (404 / DNS-fail) error, replacing the
asserted message.
- Mock _extract_via_legacy alongside _extract_via_extension in
test_network_timeout_during_extraction +
test_partial_failure_continues +
test_all_tables_fail_returns_full_failure_stats so the assertion
observes the final propagated error from the fallback chain.
- Add test_extension_per_table_failure_falls_back_to_legacy that
exercises the new behavior directly: extension scan fails with the
QueryService schema-not-authorized message
(keboola/duckdb-extension#17), legacy succeeds, parquet ends up
queryable.
The httpx client behind 'agnes query --remote' used the default 30s
timeout, killing every BigQuery SELECT that took longer than half a
minute — i.e. most non-trivial remote queries.
cli/client.py now exposes QUERY_TIMEOUT_S (default 300s, override via
AGNES_QUERY_TIMEOUT) and propagates a kw-only 'timeout' through
api_get/post/delete/patch. _query_remote passes QUERY_TIMEOUT_S so only
the long-running /api/query path gets the bump; every other CLI call
keeps the 30s default.
Server-side has no read deadline on /api/query, so the client cap was
the sole bottleneck.
Backup-orchestration commands were split across two namespaces (pull in
agnes store, push in agnes admin store), which broke the operator
mental model — pull/push are a paired operation and should sit
together.
Move pull + info into agnes admin store so all bulk operations share
one help screen. Add agnes store mine as the user-facing equivalent —
calls the same /api/store/bundle.zip endpoint with ?owner=me, which
the server resolves to the caller's user_id. Authors can archive
their own uploads without admin role; whole-Store bulk reads stay
admin-flavored as a discoverability hint.
Server: 3-line addition to export_bundle handles owner='me' as a
magic alias for the caller. No new endpoint.
Tests updated: pull/info expectations move from agnes store to
agnes admin store; new tests cover agnes store mine and the
?owner=me server resolution. 69/69 store tests green locally.
Adds whole-Store backup/restore primitives so an external CI/CD job can
mirror the Store to a git repo (and restore back from one).
REST:
- GET /api/store/bundle.zip — deterministic ZIP of all (filtered) Store
entities. Layout: manifest.json + entities/<id>/{plugin,assets}/.
Manifest carries owner_email for cross-instance restore. Auth: any
authenticated user (Store is community-open).
- POST /api/store/import-bundle — admin-only restore. Modes
merge|replace|skip; owner resolution by email with stub-disabled-user
fallback when the email is unknown on the target instance.
CLI:
- agnes store update <id> [--description X] [--zip PATH] ... — in-place
edit (server PUT permits owner OR admin per F4). Closes the missing
edit affordance for analysts who want to fix a typo or push a new
ZIP without losing install_count.
- agnes store pull [-o store.zip] [--unpack DIR] — download the bundle.
--unpack streams + extracts so an external git-backup workflow can
drop the tree straight into a repo and `git add .`.
- agnes store info [--json] — counts + size summary.
- agnes admin store push <zip-or-dir> [--mode ...] — admin-only restore.
Auto-zips a directory client-side so a working-tree → server
round-trip is one command.
cli/v2_client.py gains api_get_stream helper for binary downloads.
Tests: 5 new server tests (bundle shape + filters + round-trip + stub
user creation + skip mode + admin-only gate) + 11 new CLI tests
(update, pull/unpack, info, admin push). 66/66 store-related tests
green locally.
The previous gather used `sorted(glob, key=lambda p: p.stat().st_mtime)`.
A transient OSError (race with delete, permission flicker, EBADF on a
weird filesystem) on any single file raised through the lambda and 500-ed
the whole page.
Reworked: stat each path under try/except into a (path, stat) list, sort
the already-statted entries. Bad files drop silently from the listing.
Regression test test_profile_sessions_page_tolerates_stat_failures
patches Path.stat to raise on one of two files, asserts the page returns
200 with the good row rendered and the bad row dropped.
Devin flagged that run_session_collector still had the same audit-skip
gap I fixed in run_verification_detector and run_corporate_memory in
the previous two rounds — a PermissionError walking /home, an OSError
on /data/user_sessions mkdir, or any other unhandled exception from
collector.run() would skip the audit_log row and only show in docker
logs.
Same try/except + unhandled_error pattern as the sibling endpoints.
All three LLM-pipeline run-* endpoints now record their failures the
same way; /admin/scheduler-runs sees them. Regression test in
tests/test_admin_run_endpoints.py::TestRunSessionCollector::test_unhandled_exception_still_audits.
User feedback during e2e of #179: the listing page is nice but I want
to grab the raw jsonl and look at what's inside.
Adds GET /profile/sessions/<filename>:
- Auth via get_current_user (owner-only).
- Path safety: rejects "/", "\", "..", leading ".", and any non-".jsonl"
filename. The served path resolves under
${DATA_DIR}/user_sessions/<caller.id>/; if resolution escapes that
base directory, returns 404 (never 403, so existence of other users'
files isn't leaked).
- FileResponse with Content-Disposition: attachment.
UI: Download button per row in profile_sessions.html.
Tests in test_web_ui.py: path-traversal / nested / dotfile / non-jsonl
all 404 for owner; unauthenticated 302/401/403; authenticated owner
gets 200 + correct Content-Disposition.
Devin flagged that run_corporate_memory still had the same audit-skip
gap I just fixed in run_verification_detector — if collect_all() throws
anything other than the already-translated ValueError (DuckDB lock,
network blip, unexpected SDK error), the audit_log row was never
written and /admin/scheduler-runs missed the failure.
Same try/except + unhandled_error pattern as the verification_detector
fix from 4c4dfee8. Regression test in
tests/test_admin_run_endpoints.py::TestRunCorporateMemory::test_unhandled_exception_still_audits.
Three changes addressing user feedback during e2e test of #179 + Devin Review on e86dd5ed.
1) /profile/sessions — new self-service user page in the user menu.
Lists all session jsonls the caller uploaded via `agnes push` joined
against session_extraction_state. Each row shows uploaded_at, file
size, status badge (pending/processed/extracted), processed_at, and
items_extracted. The page docstring + help text explicitly call out
that items_extracted=0 means the verification detector ran fine but
the LLM found no claims to track — that's the documented "no items"
outcome, not a broken pipeline. Closes the gap surfaced during the
e2e test of #176 where a user could see their sessions on disk and
process them through the LLM but had no UI to inspect what happened.
2) run_verification_detector audits unhandled exceptions (Devin #1).
If detector.run() threw anything other than the already-translated
ValueError, the audit_log row was never written. The endpoint now
wraps detector.run in try/except, records the exception in
audit_params["unhandled_error"], then re-raises as 500 after audit.
The /admin/scheduler-runs page surfaces the failure row with the
error type + message.
3) SCHEDULER_AUDIT_ACTIONS list corrected (Devin #2). Previous list
had "marketplaces_sync_all" (wrong — actual is "marketplace.sync_all")
plus "data_refresh" and "scripts_run_due" which app/api/sync.py and
app/api/scripts.py don't write to audit_log. Fixed to the four
actually-logged strings; comment points at the missing audit calls
as a follow-up.
Tests: tests/test_web_ui.py adds TestAdminRoleGuards::test_profile_sessions_page_no_admin_required and tightens test_admin_scheduler_runs_page_admin_only to assert the correct marketplace.sync_all string.
create_entity + update_entity created the `scratch` temp dir inside one
try/finally but cleaned it up in a separate one. Validation HTTPExceptions
raised by _safe_zip_extract (zip_unsafe_path, zip_too_large_uncompressed)
or the BadZipFile→422 conversion exited the first scope, and the second
finally was never entered → temp dir leaked on every failed upload.
Devin flagged this on the F2 commit. The leak pre-existed (zip_unsafe_path
was the original vector); F2 added zip_too_large_uncompressed to the same
broken cleanup path. Fixed by collapsing scratch creation + cleanup into
one outer try/finally that covers both extraction AND metadata/bake; the
inner try/except/finally still handles BadZipFile→422 + tmp file cleanup.
Same restructure in update_entity. Regression test
`test_scratch_dir_cleaned_up_after_failed_extraction` triggers a
zip_unsafe_path 422 and asserts tmp/agnes_store_* contains no leaked
dirs.
Three independent reviews of PR #180 surfaced four real defects in the new
Store / my-ai-stack surface. CHANGELOG entries detail each; one-liners:
- F1 video_url XSS: any authenticated user could upload a Store entity
with `video_url=javascript:...` and pop XSS in any viewer's session via
the `<a href=...>` "Watch video" link in store_detail.html. Jinja2
autoescape doesn't block URI schemes inside attribute values. Fixed by
scheme-validating to http(s) only on create + update; 400 invalid_video_url.
- F2 ZIP decompression bomb: _safe_zip_extract checked path-traversal but
not declared file_size totals — a 50 MB compressed upload at 1:1000
ratio decompresses to 50 GB and DOS the host disk. Fixed by summing
zinfo.file_size across infolist() and refusing > 200 MB before
extractall touches disk. 413 zip_too_large_uncompressed.
- F4 admin authz parity: PUT /api/store/entities/{id} was owner-only while
DELETE allowed owner OR admin; the store-detail page hid Edit/Delete
buttons from admin even though DELETE was permitted. Fixed by allowing
admin on PUT and passing is_admin to the template; gate is now
is_owner OR is_admin everywhere.
- F5 cross-owner suffix collision: sanitize_username is many-to-one
(alice.smith / alice_smith both → alice-smith). Two such users uploading
entities with the same display name produced identical
`<name>-by-<username>` suffixes, silently colliding in the served
agnes-store-bundle on-disk paths AND the manifest catalog (Claude Code
dedupes by plugin.json `name`). Fixed by enforcing global uniqueness on
the suffixed value at create_entity; 409 conflict_global_suffix.
F3 (ZIP symlink members) was investigated and confirmed to be a
false-positive — Python's stdlib ZipFile.extractall does not honor
symlink mode bits, so no exploit exists.
9 new regression tests in tests/test_store_api.py::TestStoreSecurityFixes
covering all four. Test run locally: 60/60 store-related tests pass.
E2E test on a real BQ deploy showed every verification-extraction call
fails with HTTP 400 invalid_request_error: "output_config.format.schema:
For 'object' type, 'additionalProperties' must be explicitly set to false".
The Anthropic structured-output API now requires the field on every object
node in the json_schema. Fix: connectors/llm/anthropic_provider.py wraps
the caller-supplied schema through a recursive _strict_json_schema()
walker that adds the field where missing (preserving any explicit
override), then passes the strict variant to the API. Six unit tests in
TestStrictJsonSchema pin the recursion across nested objects, array items,
and the no-mutation invariant.
Adds /admin/scheduler-runs — a read-only admin page that surfaces the
last 200 audit-log entries from scheduler-driven actions. New
AuditRepository.query_actions(actions, limit) helper, new admin nav
entry. Failed scheduler ticks (HTTP 401, network errors) don't reach
the audit_log; the page calls that out with a hint to set
SCHEDULER_API_TOKEN if no rows show up.
The PR's #176 fail-fast change made collect_all() raise ValueError when
neither an ai: block nor ANTHROPIC_API_KEY/LLM_API_KEY was available.
verification_detector's CLI was updated to handle it; corporate_memory's
CLI was missed and crashed with an unhandled traceback.
services/corporate_memory/collector.py:main() now wraps the collect_all
call in try/except ValueError, prints a one-line actionable message
to stderr, and returns rc=1.
Regression test:
test_llm_connector.py::TestCorporateMemoryCollector::test_main_returns_1_on_no_ai_config_instead_of_traceback.
run_session_collector called collector.main() which did argparse.parse_args()
on uvicorn's sys.argv (['app.main:app', '--host', ...]) → sys.exit(2) →
SystemExit(2), which inherits from BaseException, escapes FastAPI handlers,
and propagates through the thread pool. Every scheduler tick that fired the
endpoint either 500-ed or risked killing the uvicorn worker.
services/session_collector/collector.py now exposes run(dry_run, verbose)
that returns (rc, stats); main() is a thin CLI shim that parses argv and
delegates. The admin endpoint calls run() directly and audit-logs the
per-run stats (users_processed, files_copied, files_skipped) instead of
just the rc. Three regression tests in TestRunHelper.
Closes Devin Review finding on app/api/admin.py:2819 (#179).
Devin NOTABLE: SCHEDULER_VERIFICATION_DETECTOR_INTERVAL was already
read by app/api/health.py to compute the staleness grace window, but
the actual scheduler cadence was hardcoded to 'every 15m'. The env
var name implied it controlled the cadence — it didn't. An operator
throttling the detector via the env was silently ignored by the
scheduler while the health grace silently widened.
Wired the env var into both ends. Same pattern applied to the other
two LLM-pipeline jobs:
- SCHEDULER_SESSION_COLLECTOR_INTERVAL (default 600s = 10m)
- SCHEDULER_VERIFICATION_DETECTOR_INTERVAL (default 900s = 15m)
- SCHEDULER_CORPORATE_MEMORY_INTERVAL (default 1020s = 17m)
Defaults preserve the existing 10m / 15m / 17m coprime offset so the
three jobs don't fire on the same tick.
build_jobs() now reads all three through _read_positive_int (matching
the existing pattern for data-refresh / health-check / script-runner)
and feeds them to _seconds_to_schedule. The smallest-interval check
includes the new variables so an operator can't accidentally set a
tick larger than any LLM cadence.
New tests in tests/test_scheduler.py:
- TestLLMPipelineCadenceEnvVars: env override changes the schedule
string at scheduler-init time, with parametrized invalid-value
rejection.
- TestVerificationDetectorGraceFollowsCadence: pinning the
single-source-of-truth contract — same env var moves both the
scheduler cadence and the health-check grace.
Devin BUG: /api/admin/configure seeds an ai: block to the writable
overlay at DATA_DIR/state/instance.yaml, but the three LLM consumers
imported from config.loader.load_instance_config — which reads the
static config dir only. Even if they had read the overlay, the loader
ran yaml.safe_load directly without passing through _resolve_env_refs,
so '${ANTHROPIC_API_KEY}' would have stayed a literal placeholder. The
pipeline appeared to work because the factory falls back to the env
var directly, but the overlay path itself was dead code.
Two fixes, both required:
1. Switched the three LLM consumers to app.instance_config.load_instance_config:
- services/corporate_memory/collector.py:collect_all
- services/verification_detector/__main__.py:main
- app/api/admin.py:run_verification_detector
2. app/instance_config.py runs the loaded overlay through
config.loader._resolve_env_refs *before* the deep-merge, so
'${ANTHROPIC_API_KEY}' resolves at config-load time.
New regression suite tests/test_instance_config_overlay.py pins:
- env-ref resolution against the overlay (resolved when env set,
empty when env missing — never the literal placeholder)
- deep-merge still preserves static-only sections
- the three consumers reach app.instance_config (inspected via
inspect.getsource so a future refactor that reverts the import
fails the test)
- end-to-end: a seeded overlay + ANTHROPIC_API_KEY env reaches the
factory with a resolved api_key
The PR rewrote collect_all() to call the new
create_extractor_from_env_or_config() helper, but the existing tests
still mocked the old direct create_extractor() symbol and the old
silent-skip-on-missing-config behavior. Five tests in
TestCorporateMemoryCollector and one in TestCollectorExtractorIntegration
were red on the PR branch.
Changes:
- Tests now mock connectors.llm.create_extractor_from_env_or_config
(the symbol the collector imports lazily).
- Renamed test_collect_all_no_ai_config_skips ->
test_collect_all_no_ai_config_or_env_raises and
test_collector_handles_invalid_config -> test_collector_raises_on_invalid_config.
Both assert pytest.raises(ValueError) — the explicit fail-fast
semantics defect 5 of #176 was supposed to enforce.
- collect_all() no longer swallows the factory's ValueError into
stats["errors"]; it propagates so the scheduler / admin endpoint
surface the actionable misconfiguration message instead of
pretending the run was a no-op.
- /api/admin/run-corporate-memory translates the propagated ValueError
into a 500 with the factory's message, matching
/api/admin/run-verification-detector.
Per CLAUDE.md vendor-agnostic OSS guidance — replace the real
groupon.com email used as a sanitize_username() example with a
placeholder (alice_smith@example.com).
Adds a community-driven Store where any authenticated user uploads
skills/agents/plugins as ZIPs, plus /my-ai-stack as the per-user
composition view. The served Claude Code marketplace is now:
(admin_granted ∖ opt_outs) ∪ store_installs
Skill + agent installs are merged into a single `agnes-store-bundle`
plugin in the served marketplace; type=plugin uploads stay standalone.
Names are suffixed with `-by-<owner-username>` at upload time so two
owners can use the same display name without colliding in Claude Code's
flat skill/agent namespace.
Schema v23 → v24 adds three tables:
- store_entities — community-uploaded skills/agents/plugins
- user_store_installs — what each user has chosen to install
- user_plugin_optouts — opt-out overlay on top of admin grants
Admin grant-delete drops every user's opt-out for that plugin so
re-grant resets cleanly to enabled (no sticky personal preference).
UI:
- /store — e-commerce-style listing with type/category/owner
filters, search, pagination, owner-aware [Install]
buttons, clickable cards
- /store/new — 2-step upload wizard with drag & drop, preview
validation (POST /api/store/entities/preview), docs
multi-upload, photo + video URL
- /store/{id} — detail page with hero, file list, docs, owner
actions (Edit/Delete) for the uploader
- /my-ai-stack — Granted plugins (toggle opt-out) + From the Store
(uninstall) sections
- Admin nav: Marketplaces moved into Admin dropdown, renamed to
"Curated Marketplaces"
Validation hardening: type-mismatch guards reject skill ZIP uploaded as
agent (or vice versa), and plugin ZIPs masquerading as skills/agents.
Human-readable error messages mapped client-side from machine codes.
Cross-source naming: Store entity-id-prefixed dirs (`plugins/store-<id>/`)
plus the bundle (`plugins/store-bundle/`) avoid collisions with admin
marketplaces (whose `store` slug is reserved by `is_valid_slug`).
Bundle composition is content-hashed at serve time — install/uninstall
or owner re-upload bumps the bundle's plugin.json `version`, so Claude
Code's auto-update toggle picks up changes.
Tests: 50+ new tests across naming, repositories, filter (admin ∪ store
∪ bundle), API (upload/install/uninstall/delete/preview/docs), end-to-end
marketplace.zip with bundle merging.
GET /api/health/detailed now returns a session_pipeline service entry.
Heuristic:
max(mtime of /data/user_sessions/**/*.jsonl) <=
max(processed_at in session_extraction_state) + grace_seconds
grace_seconds = 2 × verification-detector cadence (default 30 min;
configurable via SCHEDULER_VERIFICATION_DETECTOR_INTERVAL).
When the assert fails, status='warning' (never 'error') with an
actionable detail pointing at the verification-detector scheduler job.
A warning bubbles up to the existing overall='degraded' aggregation —
operators querying /api/health/detailed (or /agnes diagnose system)
get a clear breadcrumb instead of a silently-broken pipeline.
Cold-start case (no session files, or files newer than the grace
window with empty state table) is handled explicitly to avoid noise
on a fresh deploy.
Tests: tests/test_health_session_pipeline.py.
The /corporate-memory page filters status IN ('approved','mandatory')
and showed no hint that pending items exist. With approval_mode set to
'review_queue' (the default in instance.yaml.example), every collection
run would silently funnel new items into the pending bucket where no
operator ever saw them.
For admins (is_km_admin), the page now renders a banner above the
stats bar:
N pending items awaiting review — review them at /corporate-memory/admin
Non-admins see no change (the route zeroes the count server-side
before passing to the template, so the hint is never leaked).
Tests: tests/test_corporate_memory_page.py.
**BREAKING** for operators using `COMPOSE_PROFILES=full` or custom
Compose overrides that referenced these stanzas — they're gone in
docker-compose.yml and docker-compose.prod.yml. The scheduler-v2 model
(previous commit) is now the sole driver: every cadence is a job in
services/scheduler/__main__.py:JOBS hitting an admin HTTP endpoint.
Why drop instead of keep behind `profiles: [full]`:
- The previous stanzas were tight `restart: unless-stopped` boot loops.
When the scheduled run ended (every cycle), Docker re-spawned the
container, defeating any cadence the service intended.
- The whole point of #176 is that there's now exactly one driver. Two
drivers (scheduler HTTP + standalone container loop) would race on
the same /data/user_sessions and knowledge_items writes.
- Removing the stanzas is a louder signal than commenting them out —
operators upgrading get a clean failure mode (no stale containers),
not a silently double-driven pipeline.
The Python entry points (services/{corporate_memory, session_collector,
verification_detector}/__main__.py) stay — they're still callable from
the CLI for manual one-shot runs and from the new admin endpoints.
docs/architecture.md updated to reflect the new schedule table.
tests/test_docker_compose.py pins the contract: the two services must
not reappear under either Compose file.
The session-collector, verification-detector, and corporate-memory
services now run on the same scheduler-v2 model that already drives
data-refresh, health-check, script-runner, and marketplaces:
- New admin endpoints in app/api/admin.py:
POST /api/admin/run-session-collector
POST /api/admin/run-verification-detector
POST /api/admin/run-corporate-memory
All admin-gated, sync-def (FastAPI thread pool), with one audit row
per invocation. Same single-writer-of-system.duckdb pattern as the
existing /api/marketplaces/sync-all job.
- services/scheduler/__main__.py JOBS gains three entries with offset
cadences (10m / 15m / 17m, all coprime modulo the 30s tick) so the
three LLM-backed jobs don't fire on the same tick and stack their
API + DB load.
- The verification-detector endpoint surfaces the LLM factory's
fail-fast ValueError as HTTP 500 with the actionable message,
preserving the no-silent-skip contract from the previous commit.
Tests:
- tests/test_admin_run_endpoints.py covers admin gating + scheduler
registration + endpoint contract.
- tests/test_scheduler_sidecar.py existing tests continue to pass.
POST /api/admin/configure now writes a default ai: block into the
instance.yaml overlay when the request leaves it untouched and either
ANTHROPIC_API_KEY or LLM_API_KEY is set in the environment. The block
references the env var via ${VAR} syntax — secrets never land in YAML.
connectors.llm.factory grows create_extractor_from_env_or_config which
falls back to ANTHROPIC_API_KEY / LLM_API_KEY when ai_config is empty
and raises a clear ValueError when neither is available. Both
services/corporate_memory and services/verification_detector switch to
the new helper, replacing the old 'silently skip when ai: missing'
path that was the silent-failure root cause.
Tests:
- tests/test_setup_ai_block.py — overlay seeding contract.
- tests/test_llm_provider_env_fallback.py — fallback + fail-fast.
LLM provider SDKs are imported by services/corporate_memory and
services/verification_detector — both production code paths. Listing
them only in [project.optional-dependencies].dev caused the scheduler
container to boot-loop with ModuleNotFoundError on default
`docker compose up` deploys, because the Dockerfile installs core
deps only (`uv pip install --system --no-cache .`).
Adds tests/test_packaging.py to lock the contract: anthropic + openai
must live in [project].dependencies, not in dev extras.
Pre-fix: when v24 migration found rows to migrate but
data_source.bigquery.project was empty, it logged a warning per row
and returned normally. Schema_version then bumped to 24 unconditionally
→ next start's 'if current < 24:' gate skipped _v23_to_v24_finalize
forever, leaving rows in DuckDB-flavor SQL that the new
_wrap_admin_sql_for_jobs_api wrapping path rejects.
Devin escalated this from advisory ("idempotent retry") to critical
on rescan after my reply. The reply was wrong — the LIKE filter inside
the function gives idempotency IF the function is called again, but
the schema-version gate prevents that call from happening.
Fix (Devin's recommended Approach 1): raise RuntimeError BEFORE the
schema-version bump when rows need migration but project_id is empty.
The schema_version stays at 23, so on next start the 'if current < 24:'
gate fires and the migration runs again — this time with project_id
configured.
Side effect: a BQ-using deployment that hasn't set the project blocks
startup until they do. That's the right call for a config error that
would otherwise silently break all materialized tables. The error
message points at the right knob (data_source.bigquery.project +
restart).
No-rows-no-block invariant preserved: the early 'if not rows: return'
at the top of _v23_to_v24_finalize means non-BQ deployments are
unaffected.
Tests:
- test_v24_raises_when_project_not_configured_and_rows_need_migration:
asserts raise + schema_version stays at 23 (the load-bearing
invariant for retry-on-next-start to work)
- test_v24_skips_clean_when_no_rows_match_even_without_project:
asserts non-BQ deployments don't block startup
- Existing 3 tests still pass
Three items from operator feedback after running the actual flow:
(1) Help docstring lied: "--bucket / --source-table ignored" for
materialized rows. Reality: --bucket is load-bearing because
`agnes schema <name>` builds the BQ identifier as
`bq.<bucket>.<source_table>`. An empty bucket registered the row but
broke schema/describe with HTTP 400 "unsafe BQ identifier in
registry". Fix: docstring rewritten to reflect reality, plus
client-side validation rejects materialized + empty bucket with a
clear error pointing at the right knob.
(2) Post-register UX cliff: `agnes pull` after register-table reports
"Updated 0 tables (1 total)" because registration adds a registry
row but does NOT trigger a parquet build. Operators routinely
assume something's broken when they need to run
`agnes setup first-sync` to kick off the materialization. Hint
emitted on success now points at first-sync.
(3) RBAC gotcha: `agnes catalog` is RBAC-filtered via
`resource_grants`, so non-admin users don't see freshly-registered
rows until a grant is created. Hint emitted on success now points at
`agnes admin grant create <group> table <name>`.
Tests: 8/8 in test_cli_admin_materialized.py, including two new
regression tests for the validation + the hint output.
The iterative bare-name rewriter (one re.sub per name, longest-first)
was vulnerable to cross-contamination when the GCP project ID contained
a registered table name as a hyphen-delimited word.
Concrete repro:
project = 'my-ue-project'
registered = ['orders', 'ue']
user SQL = 'SELECT * FROM orders JOIN ue ON ...'
iter 1 (orders): produces 'FROM `my-ue-project.fin.orders` JOIN ue ...'
iter 2 (ue): '\bue\b' matches 'ue' INSIDE 'my-ue-project' (hyphen
creates word boundary on both sides) — corrupts
the iter-1 path
Fallback at query.py:576 caught the resulting BQ parse error and fell
back to per-table SELECT * estimate, so impact was over-estimation,
not fail-open — but the #171 partition-pruning fix silently degraded
to pre-fix behavior whenever a project name shared a hyphen-segment
with a registered table.
Fix: single re.sub call with an alternation regex sorted longest-first.
Single-pass means each source position is processed exactly once, so
freshly-inserted backticked text from one match isn't re-scanned by
later names in the alternation.
Regression test
test_rewrite_helper_does_not_corrupt_when_project_id_contains_registered_name
covers the exact Devin repro.
`_try_acquire_file_lock` opened the lock file with `open(mode='w')`
BEFORE the mtime check, which truncated the file and refreshed mtime
to now. The subsequent age check always saw ~0, so the TTL reclaim
branch was never reachable and `materialize.lock_ttl_seconds` was
a silently no-op config knob.
Repro:
before open(w): mtime age = 100000s
after open(w): mtime age = 0s
Fix: stat the lock path BEFORE any open(). If pre-probe mtime is
older than TTL, unlink (forcing a fresh inode for the open + flock
that follows). Order is now stat-then-decide-then-probe, not
probe-then-stat-then-decide.
Two regression tests added in tests/test_bq_materialize_concurrency.py:
- test_stale_held_lock_is_reclaimed_despite_live_holder — exercises
the full reclaim path with a still-living fcntl holder. Pre-fix
this returned None (in_flight forever); post-fix returns a holder
fd on a new inode.
- test_failed_probe_does_not_self_refresh_lock_mtime — sister test
pins that a failed acquisition's mode='w' truncate doesn't
pathologically loop.
Residual cross-process risk (genuinely overrunning materialize past
TTL races a fresh attempt — both write to the same parquet.tmp,
inode-level flock independence means new acquisition succeeds while
old holder is still alive) stays documented in the helper docstring.
In-process threading.Lock keyed on table_id blocks the single-process
race; cross-process protection relies on TTL being well above
longest plausible COPY (24h default).
Adds `test_unified_flow_uses_only_agnes_verbs` that asserts no `da `
substring (with trailing space, to dodge false positives on `Darwin` /
`database` / `adapter`) appears in any of the four
`resolve_lines()` shapes:
- bare (no plugins, no ca)
- plugins only
- ca only
- plugins + ca
Also pins the `agnes init --server-url … --token …` shape — commit
8784f10a's stale-on-disk-token fix relies on `init` receiving an
explicit `--token` argument; if a future refactor drops the flag from
the emitted command the test fails loudly instead of silently
regressing to 401-on-stale-token in production.
Plan: docs/superpowers/plans/2026-05-04-unified-setup-prompt.md task 10.
ZdenekSrotyr
