## Thinking Path
> - Paperclip manages agent work through governed product workflows.
> - Product E2E tests check those workflows through the real app and
runner.
> - The saved Product regression plan has 20 imported case identities.
> - The plan has no live executor, so historical baseline rows do not
qualify its current behavior.
> - The existing paid full-stack workflow already owns all 20 required
journeys.
> - This change binds TestOps selection and reporting to that workflow
without changing the product oracle or paid admission.
> - Maintainers can run the plan and inspect current case verdicts and
retained evidence.
## Linked Issues or Issue Description
**What happened?**
The standing Product regression plan has no executor. Its dashboard
points at a historical 355-result snapshot across several suites and
revisions.
**Expected behavior**
A TestOps plan run must select the intended existing full-stack cells,
execute them on GitHub, and report their original case IDs, assertions
and source revision. Missing evidence must remain a visible failure.
**Steps to reproduce**
Open the Product regression plan and inspect Executors. There is no live
job. The existing full-stack workflow accepts exact cell IDs but has no
TestOps selection or reporting path.
**Paperclip version or commit**
The branch includes reviewed master correction 04a24f4e06. This change
does not alter scenario prompts, models, deadlines or assertions.
Related public work: #18542 establishes the separate chat-catalog lane.
#18666 repairs its workflow URL binding. These changes do not implement
this Product plan.
## What Changed
- Add optional TestOps service inputs and a manual bootstrap selector to
the existing full-stack workflow.
- Map the original 20 case IDs and full names to existing native
Codex/local cells, with the existing native OpenCode/local
task-workspaces case.
- Keep reporting tokens in trusted default-branch selection and
publication jobs. Keep the existing paid provider authorization and
protected environments.
- Bind retained results to the workflow, job run, launch, source
revision and selection hash. Preserve missing-result failures and
earlier attempt verdict receipts.
- Verify uploaded identities, coverage and verdicts. Document full-plan
and subset reporting limits.
- Keep the outstanding-decision reservation live while its existing
deletion-fence test asserts the original 409 guard. Release the test
provider in finally. Production deletion and accounting code stay
unchanged.
## Verification
- `node --test tests/runner-e2e/product-regression-testops.test.mjs`: 13
positive and negative controls pass, including the separated CLI
handoff.
- Existing paid-workflow security regression passes.
- `actionlint -shellcheck= .github/workflows/runner-full-stack-e2e.yml`:
passes.
- The Product E2E catalog compatibility test confirms all 20 exact cells
exist: passes.
- `pnpm test:e2e:runner:typecheck`: passes.
- Fresh Greptile review of final head 61ef7a6e is 5/5 (11 files, no new
findings); both initial reporting findings are repaired and resolved.
- `server/src/__tests__/decision-models.test.ts`: all 37 controls pass
on a disposable official PostgreSQL 18 container in 22.02 seconds. The
unchanged focused case passed once and all five local samples; the
original Linux CI failure remains retained. This is not a claimed
deterministic local red reproduction. The new gate explicitly preserves
the outstanding-reservation precondition.
- The existing master readiness-fixture correction passes all three
retained-workspace controls on the same database in 25.79 seconds. It is
incorporated from master, not duplicated in this PR's diff.
- Final-head Linux CI also passed all 37 decision-model controls in
7.047 seconds; the server shard passed 120 files and 2,004 tests. See
[the retained successful
shard](https://github.com/paperclipai/paperclip/actions/runs/41870214517/job/126715358759).
- Previous CI failures are retained: [three stale readiness-fixture
cases](https://github.com/paperclipai/paperclip/actions/runs/41864061890/job/126694812885)
and [the reservation fixture
race](https://github.com/paperclipai/paperclip/actions/runs/41864061890/job/126694812894).
The separate [Canary Dry
Run](https://github.com/paperclipai/paperclip/actions/runs/41864061890/job/126694811605)
recorded a lost self-hosted runner connection before any step. The
fixture repairs are verified on changed source; the separate
Spot-interrupted CI attempt and its bounded retry are recorded below.
- Actual one-cell GitHub canary on the unchanged product source:
https://github.com/paperclipai/paperclip/actions/runs/41860189801 . The
native Codex/local message-marker case passed all six assertions and
cleanup in 35.24 seconds. The persisted issue is Done and its one run
succeeded. This is supporting evidence; it is not TestOps acceptance.
- Final-head CI attempt 1 was interrupted: three failed jobs have EC2
`Server.SpotInstanceTermination` receipts; six cancelled jobs show the
same simultaneous shutdown. Original logs are retained. One [failed-job
retry](https://github.com/paperclipai/paperclip/actions/runs/41870214517/attempts/2)
passed all nine interrupted checks while preserving prior successful
jobs.
- Final-head full repository CI passes, including both required `ci /
verify` and `ci / e2e` gates. GitHub still requires a non-author
code-owner approval for the workflow change; normal auto-merge is
queued. A fresh TestOps-triggered subset/full-plan acceptance remains
pending. Local projection checks do not establish a live Product pass.
## Risks
- The first bootstrap upload must establish the official TestOps job
before the saved plan can use it.
- This lane tests one declared local native configuration per case. It
does not qualify the historical 355-row matrix.
- Upload failures remain visible. A missing retained report produces
broken case results and cannot pass.
- The existing non-TestOps campaign path must keep its selection and
scheduling behavior; paid-workflow regression coverage remains in place.
## Model Used
OpenAI Codex with reasoning, code execution and tool use. The host does
not expose the exact model ID or context window to this session.
## 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
- [x] All Paperclip CI gates are green
- [x] 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>
Server tests that need a real PostgreSQL process must use
./helpers/embedded-postgres.ts instead of constructing embedded-postgres
directly.
The shared helper creates a throwaway data directory and a reserved-safe
loopback port for each test database. This protects the live Paperclip
control-plane Postgres from server vitest runs; see PAP-2033 for the incident
that introduced this guard.
