[mabry1985]

Bumps the pr-reviewer pin in the three places CI checks agree on (bundle, Dockerfile ARG, README). It also raises Vera's structural feature cap, with the budgets to match.

Why

The health alarm read: structural lane short on 12/54 rounds in 24h.

• 9 were feature-cap partials: protoAgent#4345–#4359 (25–55 features each) and campaign-plugin#39.
• 1 was a budget-timeout: #4350, which reviewed 7 of 22 features in 1500s.
• 2 had no reason recorded; pr-reviewer#282 fixes that.

v0.61.0

• pr-reviewer-plugin#282: every structural gap gets a class.
  • New relay-failure classes: no-output:max-turns and no-output. Both reason-less outages were the relay model, not protoPatch.
  • exit-4:gateway-<status> for a labelled HTTP status, and unclassified (with a structural_detail snippet) replaces other.
  • The attempt ceiling goes 3000 → 4200s.
• pr-reviewer-plugin#280: review lanes are pinned to the bound repo.

Budgets (seed; the live volume is set via POST /api/config at roll time)

key	was	now	why
structural_max_features	16	32	epic-sized PRs get full protoPatch coverage
time_budget_s	1500 (live) / 900 (seed)	3000	see the sizing below
finder_timeout_s	2100	3150	budget + 150s for the relay (p95 overhead 61s over 571 rounds); must exceed the budget so protoPatch's own SIGKILL returns the finished features as a partial
panel_attempt_timeout	3000	4050	finder + 900s tail (qaEngineer#94); under the new 4200 ceiling; the round bound is (1+1)×4050+600 = 8700s

Sizing time_budget_s:

• 86 recent protoAgent passes at structural_jobs 4 took ~353s per feature slot at p90 (~239s at p50).
• 32 features / 4 jobs = 8 slots, so ≈ 2824s at p90 (~1910s at p50).
• 3000s gives 375s per slot, the same per-slot allowance that 1500s gave 16 features.

The seed was also out of step with the live time budget (900 vs 1500). It now carries the live values.

🤖 Generated with Claude Code

[protoreview]

QA panel review — WARN

code-review-structural · head af53a289bf94 · formal

Low-risk config/version bump (plugin v0.60.0 → v0.61.0, raised panel budgets). No blockers or majors. The single surviving finding is a minor doc-drift: the new comment and README say time_budget_s "was 1500" but the seed's actual prior value was 900. Verification confirmed it as-is; nothing was refuted or re-graded. No panel disagreement, no coverage gaps, no prior requests to disposition.

Findings

	Severity	Location	Finding	Verified
🟡	minor	deploy/vera.langgraph-config.yaml:88	The new seed comment and the README state time_budget_s 'was 1500' / 'from 1500', but the seed's prior value was 900 — a doc-drift that misstates the value's o…	confirmed

findings JSON (machine-readable)

[
  {
    "file": "deploy/vera.langgraph-config.yaml",
    "line": 88,
    "severity": "minor",
    "category": "conventions",
    "claim": "The new seed comment and the README state time_budget_s 'was 1500' / 'from 1500', but the seed's prior value was 900 \u2014 a doc-drift that misstates the value's own history. Flagged by both conventions and removed-behavior review.",
    "evidence": "New comment: `#   time_budget_s         3000  (default 600; was 1500 at cap 16)` and README `time_budget_s: 3000` (from 1500); but the diff removes `-  time_budget_s: 900` and the README's prior note says the seed carried `time_budget_s: 900` since 2026-09-26. The sibling comment correctly reads `finder_timeout_s \u2026 Was 2100`, so 1500 is an outlier, not a convention.",
    "source": "panel",
    "verdict": "confirmed",
    "note": "Diff removes `time_budget_s: 900`; new comment says 'was 1500 at cap 16' and README says 'from 1500'. Prior value was 900, not 1500. Sibling `finder_timeout_s` correctly says 'Was 2100' matching the removed `finder_timeout_s: 2100`."
  }
]

[protoreview]

Promoting the WARN verdict for head af53a289bf94: all checks terminal-green, zero unresolved review threads. (approve-on-green)

Open findings carried by this approval — non-blocking, but they did not go away:

• minor deploy/vera.langgraph-config.yaml:88 — The new seed comment and the README state time_budget_s 'was 1500' / 'from 1500', but the seed's prior value was 900 — a doc-drift that misstates the value's own history. Flagged by both conventions a

Approving a WARN does not resolve its findings (issue #22).

[protoreview]

QA panel review — PASS

code-review-structural · head 21a8080464ee · formal

Clean config-only PR: plugin bump v0.60.0 → v0.61.0 and four budget increases in the seed YAML. No findings were raised, so there was nothing to verify. The round-1 minor doc-drift on time_budget_s (seed comment misstating the prior value as 1500 instead of 900) is resolved in this head — the comment and README now correctly state the prior value. No panel disagreement, no coverage gaps, no verification gaps.

No findings — the review came back clean.

findings JSON (machine-readable)

[]

---

Addressed since the last round (1 minor/nit): deploy/vera.langgraph-config.yaml:88 — no longer raised, and the PR changed the code around them since they were raised.

[protoreview]

Promoting the PASS verdict for head 21a8080464ee: all checks terminal-green, zero unresolved review threads. (approve-on-green)