zolletta-metaskill-review
读取 .zolletta-metaskill/settings.json 判定项目语言后，并行调度通用与语言特定的子工具做全量审查。把每份报告存到 .zolletta-metaskill/reports/<时间>/，并合成按功能优先（先改依赖、再按严重度）的 TODO.md，同时对比上次 TODO 验证完成情况。适合想一次性生成优先修复清单并自动校验进度时用。
▸ 展开 SKILL.md 英文原文
Full project review orchestrator. Reads the project language from .zolletta-metaskill/settings.json (written by setup), then runs the appropriate specialist skills as subagents in parallel batches — general skills (patterns, documentor) always, plus language-specific skills when applicable (e.g. python-code-style and python-testing-patterns for Python). Saves each report to .zolletta-metaskill/reports/<YYYY-MM-DD-HH-MM>/, produces an aggregated TODO.md organized by functional priority (dependency changes first, then by severity), and compares with the previous review's TODO to verify completion. Respond in the user's language.
帮我安装这个 skill：https://raw.githubusercontent.com/Zolletta/zolletta-metaskill/main/review/SKILL.mdcurl -fsSL "https://raw.githubusercontent.com/Zolletta/zolletta-metaskill/main/review/SKILL.md"# Zolletta-metaskill Review — Orchestrator You are an orchestrator that runs a full project review by invoking specialist skills in parallel batches, collecting each output, and producing an aggregated, prioritized TODO. The set of skills depends on the project's primary language: **general skills** (design patterns, documentation) always run, while **language-specific skills** run only when the project uses the matching language. ## Shared resources Read shared guidelines from the meta-skill (parent directory): - `../docs/reference/code/code-exploration.md` — code graph tools (tokensave) decision tree - `../docs/explanation/code/general-principles.md` — SOLID, KISS, composition over in