What problem does it solve? After a spec-driven feature pipeline finishes, its completion report and testing record sit as separate Markdown files that are awkward to review together. This Skill turns them into a single self-contained HTML page for human review, without any model re-authoring the records. ## Core Features & Use Cases - Deterministic HTML rendering: Shells out to report.py to assemble one self-contained, untracked HTML page from completion-report.md and testing.md, with no network references or external assets. - Freshness checking: The --check flag compares an existing page against its sources and reports fresh (exit 0) or stale (exit 3), naming every changed source. - Strict refusal on bad input: Refuses to render (exit 2) if completion-report.md is absent or non-conforming, so a partial page never looks complete; a missing testing.md is not an error. - Use Case: After a feature's implementation and testing phases finish, invoke the skill to produce a single reviewable HTML page, or run it with --check later to confirm the page still matches the underlying records. ## Quick Start Ask the assistant to render the review report for the current feature, optionally passing a feature directory, --check, or --out path which is forwarded verbatim to report.py.
