Make no-reviewer review blocks stop saying advisory-only #790

New issue

Closed

opened 2026-07-09 09:06:10 +00:00 by coilyco-ops · 2 comments

coilyco-ops commented

2026-07-09 09:06:10 +00:00

Member

Problem

A no-reviewer in-container review result is converted to a fail-closed block, but the human-facing text still says the opposite:

review note: ADVISORY-ONLY REVIEW: no reviewer family was available for codex, so the adversarial panel could not run and did NOT gate this diff.
panel BLOCKED: 0/0 passing - do NOT land this diff.
review summary: blocked: ADVISORY-ONLY REVIEW: no reviewer family was available for codex, so the adversarial panel could not run and did NOT gat...
ward: review gate BLOCKED: 0/0 passing (class default); blocked: ADVISORY-ONLY REVIEW: no reviewer family was available for codex, so the adversarial panel could not run and did NOT gat...; landing refused

The gate behavior is correct: no runnable reviewer must fail closed. The regression is the contradictory wording and summary/handoff shape. It tells the worker and human reader that the panel was advisory-only and did not gate the diff, then immediately blocks landing.

Evidence

The pasted run was a Codex worker in /workspace/factory-game-v3 at 2026-07-09 01:48:00.125. All reviewer families were unavailable:

codex launch probe failed after printing the Codex startup banner and prompt text.
claude smoke test produced no usable output.
opencode was not on PATH.
goose could not reach the Ollama endpoint.

/substrate/ward shows the source of the contradiction:

internal/reviewpanel/panel.go renders advisoryNote as ADVISORY-ONLY REVIEW ... did NOT gate this diff.
cmd/ward/agent_review.go then converts GateAdvisory to GateBlock before writing the summary, handoff, panel output, and issue conclusion.
reviewSummaryBlocked prefixes that stale advisory note with blocked:.

Expected behavior

When no reviewer can run and ward converts the result to a fail-closed block, every user-visible surface should say the gate blocked because no reviewer was runnable. It should not use ADVISORY-ONLY REVIEW, and it should not say the diff was not gated.

Surfaces to cover:

stderr panel note
panel BLOCKED text
review summary
~/.ward/review-summary.txt
WARD-REVIEW / blocked error path
issue conclusion comment body

Do

Preserve fail-closed behavior for no runnable reviewers.
Change the note/summary wording after advisory-to-block conversion so the unavailable panel reads as a block, not an advisory bypass.
Add regression coverage for the no-runnable-reviewer path. The test should fail if a blocked result contains ADVISORY-ONLY REVIEW or did NOT gate this diff in the worker-facing summary/output.
Update docs/dispatch-review.md only if the public contract needs clarifying.

Acceptance

A no-runnable-reviewer review still exits non-zero and refuses landing.
The output and summary clearly say the review gate blocked fail-closed because no reviewer family was available.
No blocked no-reviewer summary says ADVISORY-ONLY REVIEW or did NOT gate this diff.
ward exec test passes.

Dispatch note

Dispatch this with --skip-review because the issue repairs review-gate output and should not be blocked by the broken review gate it is fixing. This follow-up narrows the remaining regression after ward#784.

## Problem A no-reviewer in-container review result is converted to a fail-closed block, but the human-facing text still says the opposite: ```text review note: ADVISORY-ONLY REVIEW: no reviewer family was available for codex, so the adversarial panel could not run and did NOT gate this diff. panel BLOCKED: 0/0 passing - do NOT land this diff. review summary: blocked: ADVISORY-ONLY REVIEW: no reviewer family was available for codex, so the adversarial panel could not run and did NOT gat... ward: review gate BLOCKED: 0/0 passing (class default); blocked: ADVISORY-ONLY REVIEW: no reviewer family was available for codex, so the adversarial panel could not run and did NOT gat...; landing refused ``` The gate behavior is correct: no runnable reviewer must fail closed. The regression is the contradictory wording and summary/handoff shape. It tells the worker and human reader that the panel was advisory-only and did not gate the diff, then immediately blocks landing. ## Evidence The pasted run was a Codex worker in `/workspace/factory-game-v3` at `2026-07-09 01:48:00.125`. All reviewer families were unavailable: * codex launch probe failed after printing the Codex startup banner and prompt text. * claude smoke test produced no usable output. * opencode was not on `PATH`. * goose could not reach the Ollama endpoint. `/substrate/ward` shows the source of the contradiction: * `internal/reviewpanel/panel.go` renders `advisoryNote` as `ADVISORY-ONLY REVIEW ... did NOT gate this diff`. * `cmd/ward/agent_review.go` then converts `GateAdvisory` to `GateBlock` before writing the summary, handoff, panel output, and issue conclusion. * `reviewSummaryBlocked` prefixes that stale advisory note with `blocked:`. ## Expected behavior When no reviewer can run and ward converts the result to a fail-closed block, every user-visible surface should say the gate blocked because no reviewer was runnable. It should not use `ADVISORY-ONLY REVIEW`, and it should not say the diff was not gated. Surfaces to cover: * stderr panel note * `panel BLOCKED` text * `review summary` * `~/.ward/review-summary.txt` * `WARD-REVIEW` / blocked error path * issue conclusion comment body ## Do * Preserve fail-closed behavior for no runnable reviewers. * Change the note/summary wording after advisory-to-block conversion so the unavailable panel reads as a block, not an advisory bypass. * Add regression coverage for the no-runnable-reviewer path. The test should fail if a blocked result contains `ADVISORY-ONLY REVIEW` or `did NOT gate this diff` in the worker-facing summary/output. * Update `docs/dispatch-review.md` only if the public contract needs clarifying. ## Acceptance * A no-runnable-reviewer review still exits non-zero and refuses landing. * The output and summary clearly say the review gate blocked fail-closed because no reviewer family was available. * No blocked no-reviewer summary says `ADVISORY-ONLY REVIEW` or `did NOT gate this diff`. * `ward exec test` passes. ## Dispatch note Dispatch this with `--skip-review` because the issue repairs review-gate output and should not be blocked by the broken review gate it is fixing. This follow-up narrows the remaining regression after ward#784.

coilyco-ops commented

2026-07-09 09:06:20 +00:00

Author

Member

🔒 Reserved by ward agent --harness codex — container engineer-codex-ward-790 on host kais-macbook-pro-2.local is carrying this issue (reserved 2026-07-09T09:06:18Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (1h TTL); --force overrides.

Do not comment on or edit this issue to steer the run while it is reserved. The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a new issue, dispatched fresh — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).

run seed context — what this run is carrying (ward#609)

Resolved: coilyco-flight-deck/ward#790 · branch issue-790 · harness codex · workflow pr
Run: engineer-codex-ward-790 · ward v0.488.0 · dispatched 2026-07-09T09:06:18Z
Comment thread: 0 included in the pre-flight read, 0 stripped (ward's own automated comments).

Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.488.0).

— Codex, via ward agent

🔒 Reserved by `ward agent --harness codex` — container `engineer-codex-ward-790` on host `kais-macbook-pro-2.local` is carrying this issue (reserved 2026-07-09T09:06:18Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (1h TTL); `--force` overrides. **Do not comment on or edit this issue to steer the run while it is reserved.** The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a **new issue, dispatched fresh** — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494). <details><summary>run seed context — what this run is carrying (ward#609)</summary> - **Resolved:** `coilyco-flight-deck/ward#790` · branch `issue-790` · harness `codex` · workflow `pr` - **Run:** `engineer-codex-ward-790` · ward `v0.488.0` · dispatched `2026-07-09T09:06:18Z` - **Comment thread:** 0 included in the pre-flight read, 0 stripped (ward's own automated comments). Static container doctrine and seed boilerplate are identical every run and omitted here (they ride ward v0.488.0). </details>  — Codex, via `ward agent`

coilyco-ops referenced this issue from a pull request that will close it,

2026-07-09 09:11:40 +00:00

Make no-reviewer review blocks stop saying advisory-only #794

coilyco-ops commented

2026-07-09 09:13:14 +00:00

Author

Member

WARD-OUTCOME: done - rewrote no-reviewer block wording and covered the fail-closed surfaces
workflow: pr; review summary: skipped by --skip-review / --no-review-gate

The conversion seam was the right place to fix this. Once I rewrote the advisory fallback at the block boundary, the summary, handoff, panel output, and issue comment all lined up. The only friction was Forgejo PR creation and status polling, which I handled through the API and tasks endpoint. Confidence is high.

WARD-OUTCOME: done - rewrote no-reviewer block wording and covered the fail-closed surfaces workflow: pr; review summary: skipped by --skip-review / --no-review-gate The conversion seam was the right place to fix this. Once I rewrote the advisory fallback at the block boundary, the summary, handoff, panel output, and issue comment all lined up. The only friction was Forgejo PR creation and status polling, which I handled through the API and tasks endpoint. Confidence is high.