agenvoy/skill-code-reviewer

Última actividad de origen registrada: 21 de septiembre de 2026 a las 15:45

Catálogo de SkillsMP actualizado: 2 de noviembre de 2026 a las 18:02

skills recopiladas: 1

Estrellas en GitHub: 1
Forks en GitHub: 0

Instalar con un asistente de IA

Copia este prompt en el asistente de IA que estés usando.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/agenvoy/skill-code-reviewer.

README del repositorio

Archivo de origen del README: README.md

Leer el README original

Mostrar README completo

> [!NOTE] > This README was generated by [SKILL](https://github.com/agenvoy/skill-readme-generate), get the ZH version from [here](./doc/README.zh.md). > This skill's implementation was fully agent-generated; the developer only tuned its input/output. *** AST-DRIVEN CODE REVIEWS THAT SKIP THE NOISE! <a href="LICENSE"><img src="https://img.shields.io/github/license/agenvoy/skill-code-reviewer?include_prereleases&style=for-the-badge" alt="License"></a> *** > An agent skill with AST-based multi-language analysis, per-finding source validation, and project convention checks ## Table of Contents - [Features](#features) - [Architecture](#architecture) - [License](#license) ## Features > `/code-reviewer [PROJECT_PATH] [OUTPUT_FILE]` · [Documentation](./doc/doc.md) - **AST-Driven Multi-Language Analysis** — Go runs through a `go/ast` helper that attaches source snippets, Python uses the built-in `ast`, and JS/TS gets a structural scan plus the project's eslint, each degrading to string scanning when its toolchain is missing. - **Per-Finding Validation Pass** — Every issue and suggestion is re-checked against the source before it reaches the report, and anything unconfirmed is dropped rather than downgraded so false positives never erode trust in the rest. - **Project Convention Adherence** — Code is checked against the `CLAUDE.md` / `AGENTS.md` files in its own directory and parents, each violation quotes the rule verbatim, and trade-offs already marked with `nolint` or a comment are skipped. - **Entropy-Based Secret Detection** — Beyond keywords, suspicious strings are scored by Shannon entropy and filtered against UUID, hash, and MIME-type patterns to cut false positives in credential detection. - **No Report Without Findings** — With zero issues and no actionable suggestions, it creates no directory and writes no file, and suggestions themselves are barred from speculative or decorative content. ## Architecture > [Full Architecture](./doc/architecture.md) ```mermaid graph TB User[User] -->|/code-reviewer| SKILL[SKILL.md Orchestration] SKILL --> Analyze[analyze_code.py Detect and Dispatch] Analyze --> Lang[Go / Python / JS-TS Analyzers] Lang --> JSON[issues / metrics JSON] JSON --> Validate[Validation Pass + Convention Check] Validate --> Gate{No-Op?} Gate -->|Yes| Msg[One-line no-op message] Gate -->|No| Report[.doc/code-reviewer/ report] ``` ## License This project is licensed under the [MIT LICENSE](LICENSE).

Cargando el README del repositorio…

Skills en este repositorio

clasificación pendiente

skill

ocupación

descripción

actualizado

code-reviewer

sin clasificar

Analyze project source code and generate optimization suggestions. Use when user wants code review, performance optimization advice, security hardening recommendations, or architecture improvement suggestions.

Idioma del texto original: chino

2026-09-21

Mostrando 1 de 1 skills recopiladas.