review-readiness-cleanup

name	review-readiness-cleanup
description	Inspect review packets, review plans, finding registers, demo or proof registers, and milestone review docs for structural readiness, classifying safe cleanup, blockers, skipped surfaces, and follow-on needs without remediating findings, publishing reports, or approving review readiness.

Review Readiness Cleanup

Inspect one review packet or milestone review surface before an internal, external, or CodeBuddy-style review cycle starts. This skill reduces avoidable review friction from stale markers, missing metadata, unclear blockers, and packet-structure drift.

It is a readiness cleanup classifier, not a qualitative reviewer and not an approval authority.

Quick Start

1. Confirm the review root and intended review cycle.
2. Confirm the cleanup policy stops before remediation, publication, and review approval.
3. Run the deterministic helper when local filesystem access is available:
   • scripts/inspect_review_readiness.py --review-root <path> --out <artifact-root> --run-id <run_id>
4. Review the Markdown and JSON artifacts.
5. Route actionable results to the appropriate owner. Do not rewrite findings, alter severity, publish reports, or approve review readiness from this skill.

Required Inputs

At minimum, gather:

• mode
• review_root
• target
• policy

Supported modes:

• inspect_review_packet
• inspect_milestone_review
• refresh_readiness_cleanup

Useful policy fields:

• write_cleanup_artifact
• allow_safe_mechanical_cleanup
• require_finding_register
• require_demo_or_proof_register
• stop_before_remediation
• stop_before_publication
• stop_before_review_approval

If the review root is missing or intentionally not in scope, return skipped with the reason visible.

Classification Model

Classify findings as:

• safe_mechanical_cleanup: stale placeholders, unchecked metadata, empty packet sections, or formatting drift that can be repaired without changing substantive findings.
• blocker: missing required review surfaces, explicit blocker markers, unresolved high-priority review state, or evidence drift that would make the review misleading.
• skipped: a review surface was intentionally absent, unavailable, or out of scope for this cycle.
• follow_on_needed: useful cleanup or process improvement that should not block the current review but should be queued.

If evidence is insufficient, classify as blocker or skipped; do not infer readiness from absence.

Output

Write Markdown and JSON artifacts when an output root is available.

Default artifact root:

.adl/reviews/review-readiness-cleanup/<run_id>/

Required artifacts:

• review_readiness_cleanup_report.md
• review_readiness_cleanup_report.json

Use the detailed contract in references/output-contract.md.

Stop Boundary

This skill must not:

• run a full internal or external review
• remediate review findings
• rewrite finding severity, disagreement, evidence, or conclusions
• publish customer-facing reports
• approve review readiness without evidence
• create issues or PRs without explicit operator approval

Handoff candidates:

• documentation-specialist for approved mechanical doc cleanup.
• gap-analysis when expected review scope needs comparison against evidence.
• finding-to-issue-planner when cleanup or blockers should become issue candidates.
• review-quality-evaluator when the review packet itself needs a quality gate.