Run Reviews

What it does

Triggers independent multi-axis review across seven specialist reviewer agents — governance, domain-validator, architecture, performance, standards, observability, and security — plus an independent GR deep-review engine. Each reviewer is independent of the developer and of each other. A BLOCK from any confirmed finding halts the pipeline.

When to use it

Run after all test tiers are GREEN. Reviewers are read-only; they consume the dev-reports, test reports, and the actual code and emit verdict reports.

Inputs

• ›REQ id
• ›TASK-N-dev-report.md files
• ›TASK-N-test-*-report.md files

Outputs

• ›TASK-N-review-governance.md
• ›TASK-N-review-architecture.md
• ›TASK-N-review-performance.md
• ›TASK-N-review-standards.md
• ›TASK-N-review-observability.md
• ›TASK-N-review-security.md
• ›TASK-N-review-domain-validator.md
• ›gr-review.md

Example

/eng-org:run-reviews REQ-20260723-d904-01 — spawns seven reviewer agents and the GR engine in parallel; each writes a verdict of APPROVE, NEEDS-CHANGES, or BLOCK; any BLOCK returns to the TL for a fix iteration.