plugin-review

<!--

Review a Claude Code plugin for principles alignment, terminal UX quality, and documentation freshness via orchestrator–subagent analysis.

<!--

Fix-forward agent for build and test failures. Receives failing test output and recently-modified files, implements the minimum fix to restore tests to passing. Called by the orchestrator in Phase 4.5 when run-build-test.sh reports failures.

Track C analysis — reads plugin documentation against implementation structure and returns per-file freshness assessment with drift classification.

Track D analysis — reads plugin implementation files and returns per-principle context efficiency status table with enforcement assessment and JSON assertions.

Targeted implementation agent for assertion-driven fixes. Receives a list of failing assertions with context and implements the minimal fix for each. Called by the orchestrator after run-assertions.sh finds failures.

Track A analysis — reads plugin implementation files and returns per-principle status table with enforcement layer assessment and root architectural alignment.

Receives a list of previously-fixed findings and the current state of relevant files, then performs a narrative re-check to verify each fix is still intact. Returns per-finding holding/regressed status with evidence.

Track B analysis — reads user-facing code paths and returns severity-grouped terminal UX findings across information density, user input, progress/feedback, and terminal-specific criteria.

<!--

<!--

<!--

<!-- architectural-context

2 hooks across 1 event