2 Commits
Author SHA1 Message Date
Devin FoleyandPaperclip 6d117998c8 fix(ci): skip reviewed-lock checks for a pending Dependabot lock in the Runner support suite (#21942)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work
> - Dependabot pull requests that change `pnpm-lock.yaml` are the only
way, other than the lockfile refresh, that lock content changes
> - #19356 let a verified Dependabot lockfile pass CI. The server's
reporting-lock policy test skips when `PAPERCLIP_CI_PR_AUTHOR` is
`dependabot[bot]`, and the lockfile refresh renews the proof after the
merge
> - #19537 then made the Runner support suite a required check. That job
does not set `PAPERCLIP_CI_PR_AUTHOR`, and the suite's own reviewed-lock
checks have no Dependabot skip
> - So every Dependabot lockfile pull request fails `ci / Runner support
suite` and `ci / verify`. The last one that merged, #17364, merged
before #19537
> - This pull request gives the support suite the same Dependabot skip
as the server test
> - The benefit is that Dependabot dependency updates can merge again,
with the same lock trust rules as before

## Linked Issues or Issue Description

Refs #19356, #19537, #19614, #17172, #19575. Rebased onto `master` after
#21943, which fixed the unrelated `qualify_test_selection` failure that
a CI file change triggers.

**What happened?**

Dependabot pull request #17172 changes `pnpm-lock.yaml`. `ci / Runner
support suite` fails in four test files, and `ci / verify` fails because
of it:

- `prepare-report-lock.test.ts` > installs only the reviewed dependency
correction (4 cases): the committed lock hash is not in
`report-lock.json`.
- `source-sdk-public-type.test.ts`: the shared setup throws "Committed
SDK reference lock is not reviewed".
- `source-sdk-evidence.test.ts` > seals a normal producer-certified
reviewed-reference SDK: the same error.
- `source-sdk-public-inputs.test.mjs` > the committed SDK public source
pins match this checkout's committed tree: `pnpm-lock.yaml` is stale in
`source-sdk-public-inputs.json`.

A Dependabot pull request may not change `report-lock.json`,
`report-lock.patch` or the SDK pins. `lockfile-trust.mjs` rejects that
change, and `refresh-lockfile.yml` renews those files after the merge.
So the pull request can never pass.

**Expected behavior**

A verified Dependabot lockfile passes the Runner support suite, as it
already passes the server's reporting-lock policy test. Other authors
keep every check.

**Steps to reproduce**

1. Open a Dependabot pull request that changes `pnpm-lock.yaml`, such as
#17172.
2. Let CI run. `ci / Runner support suite` fails in the four files
above.

## What Changed

- `.github/workflows/pr-trusted.yml`: the `verify_runner_e2e_support`
job sets `PAPERCLIP_CI_PR_AUTHOR` from the pull request author, like the
general test job.
- `tests/runner-e2e/prepare-report-lock.test.ts`,
`source-sdk-evidence.test.ts`, `source-sdk-public-inputs.test.mjs`: the
test cases that need a reviewed committed lock skip when
`PAPERCLIP_CI_PR_AUTHOR` is `dependabot[bot]`. This uses the same check
as `server/scripts/reporting-lock-policy.test.mjs`. The other cases in
these files still run.
- `tests/runner-e2e/vitest.config.ts`: on a Dependabot pull request, the
suite excludes `source-sdk-public-type.test.ts`. Its shared `beforeAll`
prepares the reviewed SDK reference, so every case in the file needs a
reviewed lock.
- `tests/runner-e2e/source-sdk-public-type.ts`: the installed
TypeScript, Babel parser and Zod package hash checks move into an
exported `verifySdkProofPackages`, which `verifySdkPublicSource` calls.
The lockfile refresh renews the lock and source pins after a merge, but
keeps these package hashes.
- `tests/runner-e2e/source-sdk-proof-packages.test.ts` (new): runs
`verifySdkProofPackages` against the installed packages on every pull
request, including Dependabot ones, and checks that a changed package
pin is rejected. So a Dependabot update of one of those packages still
fails until its pins are reviewed.

## Verification

- Red, then green, with a real Dependabot lock: I committed the
`pnpm-lock.yaml` and `package.json` changes of #17172 on top of this
branch in a local commit (not pushed). Then I ran `node
node_modules/vitest/vitest.mjs run --config
tests/runner-e2e/vitest.config.ts` on the four files:
- Without `PAPERCLIP_CI_PR_AUTHOR`: 4 files failed and the same 6 tests
failed as in CI.
- With `PAPERCLIP_CI_PR_AUTHOR=dependabot[bot]`: 3 files passed, 33
tests passed and 6 skipped. `source-sdk-public-type.test.ts` was
excluded.
- On this branch with master's lock: `pnpm test:e2e:runner:unit` passed
(Vitest: 687 files, 11640 tests; node tests passed).
- `node --test .github/scripts/tests/*.test.mjs`: 808 passed.
- After the package check change: the five affected files
(`source-sdk-proof-packages`, `source-sdk-public-type`,
`source-sdk-evidence`, `source-sdk-public-inputs`,
`prepare-report-lock`) pass, 71 tests. `tsc --noEmit -p
tests/runner-e2e/tsconfig.json` passes.
- `pr.yml` calls `pr-trusted.yml@master`, so this pull request's own CI
does not use the new job variable. The first Dependabot pull request
after the merge checks it.

## Risks

- A Dependabot update of TypeScript, the Babel parser or Zod still fails
the suite until a maintainer reviews the new package pins. That is
deliberate.
- On a Dependabot pull request, the suite no longer checks the committed
lock against the reviewed proof. The policy job still requires that the
lock comes from a verified Dependabot commit (`lockfile-trust.mjs pr`),
and the lockfile refresh still admits the lock and renews the proof and
the SDK pins after the merge. That is the same trust as #19356.
- `PAPERCLIP_CI_PR_AUTHOR` is the pull request author from the GitHub
event. A contributor cannot set it to `dependabot[bot]`. Other commits
pushed to a Dependabot branch are still checked by the policy job.
- No runtime code changes.

## Model Used

- Claude Opus 5.5 (`claude-opus-5-5`), Anthropic, used through Claude
Code with tool use and code execution.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-11-16 18:03:41 -08:00
f4337bb5fb test(evals): land the chat catalog eval harness on master (#19614)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work
> - The chat-feature evals measure whether agents complete real chat
requests through the product
> - The chat catalog eval harness and its repairs live on the
`chat-eval/master-candidate` branch, not on master
> - Today's work made 139 of 242 chat evals pass in verified single-case
live runs. Each product fix landed on master as its own pull request
> - The eval-side changes (setups, graders, fault seams, evidence
capture) still exist only on the candidate
> - This pull request lands the harness on master. Master stays the
source of truth for product behavior
> - The benefit is one tree: evals run against master commits, and
product fixes reach the evals as soon as they merge

## Linked Issues or Issue Description

Refs #19441, Refs #19599, Refs #20270, Refs #20334, Refs #20342, Refs
#20344, Refs #20354, Refs #20360, Refs #20373, Refs #20382, Refs #20409

**What existing behavior does this improve?**
The chat catalog eval harness and the product it evaluates live on a
long-running branch outside master. Eval fixes do not reach master, so
master-based eval runs use old setups and graders.

**Subsystem affected**
`tests/runner-e2e` chat catalog harness, plus a small set of product
changes that exist only on the candidate.

**Proposed change**
Land the harness on master. Every product fix from this eval work
already merged to master as its own pull request. Conflicts keep
master's product behavior.

## What Changed

- Most of the diff (808 of 851 files) is under `tests/runner-e2e/`:
chat-catalog domains, scenarios, descriptors, the TestOps manifest,
private evidence tooling, and their tests.
- Graders now check saved state, setup literals and rendered or clicked
links. Many keyword and regex checks on free-form replies are removed.
Each changed grader has a bumped `oracleVersion` and calibration tests
(a valid reply, a reworded valid reply, and broken variants that must
fail).
- The shared link helper `hasLink` no longer treats a list link (for
example `/COM/issues?view=mine`) as a link to every task without a URL
key.
- Setup fixes build real state through supported product routes: routine
history, agent grants through the board grant route, budget holds with a
real cap, and a storage-layer archive fault for
`inbox-search.archive-partial-failure`.
- Browser boundary journeys (sign-in, chat-disabled, onboarding) now
write the standard pass evidence files.
- Retargeted cases, approved by the maintainer:
`agent-chat.chat-disabled` (Conference Room gate),
`agents.restore-instructions` (no saved instruction history),
`access.manage-membership` (member permission edits stay board-only).
- The private evidence trusted files (`manual-private-handoff.ts`,
`manual-private-retention.ts`, `manual-private-report-assets.ts`) are
byte-identical to master.
- Product files outside `tests/` (43 files, +1,328 −49) carry
candidate-only changes that master does not yet have. Most are tests.
The main code changes are in `request-human-input.ts`,
`heartbeat-run-summary.ts`, `native-completion-feedback.ts`,
`native-run-finalizer.ts`, `openapi.ts`, `secrets.ts` and
`IssueChatLongThreadPerf.tsx`.

## Verification

- `pnpm vitest run chat-catalog-domains chat-catalog-contract.test.ts
chat-catalog-consistency.test.ts unit-launcher.test.mjs
--testTimeout=180000` in `tests/runner-e2e`: 269 files pass, 3,550 tests
pass.
- `node --test chat-full-catalog-testops-manifest.test.mjs`: 10 pass.
- `node tests/runner-e2e/source-sdk-public-inputs.mjs <HEAD> --check`:
no change.
- `tsc` for `server` and `tests/runner-e2e`: clean.
- Live single-case verification: 139 of 242 cases passed on paid runs,
and a person read the saved evidence for each. These are single-case
runs, not one full-cohort run.

## Risks

- Large diff. Most of it is test-only, but the 43 non-test files change
product code paths. Review those files first.
- Some graders are now less strict on reply wording. A person must read
replies for meaning on those cases.
- 42 cases need disposable outside accounts (Notion, Slack, GitHub,
Daytona and others) and stay blocked. `access.invite-teammate` stays
failing until local agent isolation lands, because the runner's durable
state keeps a live invite token.
- The paid workflow compares the trusted private-evidence files with
master. They are unchanged, so this risk is low.

## Model Used

Anthropic Claude Opus 5.5 (`claude-opus-5-5`), with tool use and code
execution, in Claude Code. Subagents used the same model.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-11-10 08:02:11 -06:00