[deep-report] Fix PR-review cluster trigger gates — 272 wasted run-attempts/day across 8 agents

[github-actions]

Context

Identified by DeepReport intelligence briefing 2026-05-12 (Agent Performance Report #31690).

The PR-review cluster of 8 agents (Q, Scout, Archie, /cloclo, Grumpy, Security Review, PR Nitpick, PR Code Quality) shows: Quality ~20/100, Effectiveness ~10/100. Out of 41 runs in the sample window: 33 skipped, 8 action_required, 0 successful outputs. This is ~272 wasted run-attempts/day — the highest waste in the ecosystem.

Root cause: trigger gates not filtering properly; agents spin up for irrelevant PRs.

Expected Impact

• Reclaim ~272 wasted run-attempts/day (highest impact in fleet)
• Reduce GitHub Actions minutes and engine token spend
• Reduce noise in daily audit reports

Suggested Approach

• Audit each PR-review workflow's on.pull_request / on.labels triggers
• Add filters: paths, labels, types: [labeled, ready_for_review], or if: conditions checking author or branch name
• Consider consolidating overlapping reviewer agents (8 agents reviewing every PR is excessive)

Suggested Agent

Agentic Maintenance or a coding agent — touches ~8 workflow files.

Estimated Effort

2-3 hours of agent time.

---

Source: DeepReport 2026-05-12 #aw_dr2

Generated by DeepReport - Intelligence Gathering Agent · ● 9.5M · ◷

• expires on May 14, 2026, 3:42 PM UTC