Example Review Report

This is a real output from /zolletta-metaskill review run on the zolletta-metaskill project itself (dogfooding). The report was generated on 2026-08-06 and is preserved here as a reference for what the review output looks like.

Files

File	Description
SUMMARY.md	Executive summary with overall grade, grades by area, strengths, weaknesses, and trend vs previous review
TODO.md	Prioritized action items (critical, high, medium, low) with source, files, problem, and fix
patterns.md	Design pattern review — God classes, SOLID, coupling, composition vs inheritance
documentor.md	Documentation review — Diátaxis compliance, drift detection, broken links, API validation
python-code-style.md	Python source code style — ruff, mypy, naming, docstrings, type annotations
python-testing-style.md	Python test code review — isolation, naming, coverage gaps, mocking, fixtures

Report structure

Every review produces a timestamped run folder under .zolletta-metaskill/<YYYY-MM-DD-HH-MM>/ containing reports/ (LLM judgment) and cache/ (deterministic script outputs). The orchestrator launches subagents for each review area in parallel, each writing its own report to reports/. The orchestrator then aggregates grades into reports/SUMMARY.md and collects actionable items into reports/TODO.md.

See Reports for the file format specification and templates.