Token导航 LogoToken导航

Agent Skill · 研究检索

platonic-code-review(柏拉图式的代码审查)

platonic-code-review 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

platonic-code-review 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 58,684。

总安装 408周安装 17GitHub Stars 3下载 136核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:platonic-code-review(柏拉图式的代码审查)
来源仓库:https://github.com/caesar0301/platonic-coding-skills
仓库路径:skills/platonic-code-review
安装命令:
npx skills add https://github.com/caesar0301/platonic-coding-skills --skill platonic-code-review
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/caesar0301/platonic-coding-skills --skill platonic-code-review

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息,适合快速定位候选结果。适用于关键词搜索、任务场景匹配和来源线索筛选等研究检索场景。通过 npx skills add 命令从 GitHub 仓库安装,需结合原始 README 核验用法。安装前建议确认权限范围、维护状态及是否会触发联网或文件读写。

可配合来源仓库进一步了解技能的实际功能和限制条件。

来源仓库内路径:skills/platonic-code-review

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/caesar0301/platonic-coding-skills --skill plat…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

platonic-code-review 怎么安装?

在终端执行 npx skills add https://github.com/caesar0301/platonic-coding-skills --skill platonic-code-…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

platonic-code-review 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

platonic-code-review 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

platonic-code-review 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

platonic-code-review 用的人多吗?

公开数据是累计安装 408 次,近一周安装 17 次,下载 136 次,GitHub 3 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

platonic-code-review 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
platonic-code-review

柏拉图式的代码审查 · 本页

4083需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Platonic Code Review

Validate that code implementation is consistent with project specifications and requirements.

When to Use This Skill

Use this skill when you need to:

  • Verify code implementation matches RFC specifications
  • Validate feature completeness against requirements
  • Check consistency between specs and actual code
  • Identify gaps where specs exist but implementation is missing
  • Find discrepancies where code behavior differs from specs
  • Audit compliance with documented design decisions

Keywords: spec compliance, implementation validation, requirement verification, consistency check, gap analysis

What This Skill Does

This is NOT a general code quality reviewer. It specifically focuses on:

✅ Spec-to-Code Consistency: Does the code implement what the specs describe? ✅ Completeness: Are all specified features implemented? ✅ Correctness: Does the implementation match spec requirements? ✅ Gap Identification: What's specified but not implemented? ✅ Discrepancy Detection: Where does code differ from specs?

Default Behavior

By Default: Generates review report without modifying code

  • Creates detailed report of findings
  • Highlights inconsistencies and gaps
  • Provides actionable recommendations
  • Does NOT make code changes automatically

To Modify Code: User must explicitly request code modifications

Review Procedure

The standard review process follows these steps:

Step 1: Understand Specifications

  • Locate and read relevant specification documents

- RFC documents (RFC-NNNN.md convention) - Requirements documents - Design documents - API specifications

  • Extract key requirements and expected behaviors
  • Identify testable criteria

Step 2: Generate Functionality Checklist

Create a comprehensive checklist from specs:

  • List all specified features
  • List all specified behaviors
  • List all API endpoints/interfaces
  • List all data structures
  • List all business rules
  • List all constraints and validations

Step 3: Map Specs to Code

Identify code locations implementing each spec:

  • Find relevant files and modules
  • Map features to implementation
  • Identify entry points and key functions
  • Understand code structure

Step 4: Review Implementation Against Specs

For each checklist item:

  • Status: ✅ Implemented / ⚠️ Partial / ❌ Missing / 🔍 Unclear
  • Consistency: Does implementation match spec?
  • Completeness: Are all aspects implemented?
  • Correctness: Does it work as specified?
  • Evidence: Specific code references

Step 5: Identify Discrepancies

Document inconsistencies:

  • Implementation differs from spec
  • Missing features specified but not implemented
  • Extra features implemented but not specified
  • Incorrect behavior doesn't match spec requirements
  • Incomplete implementation partially done

Step 6: Generate Review Report

Create structured report with:

  • Executive summary
  • Functionality checklist with status
  • Detailed findings (consistent, inconsistent, missing)
  • Code references for each finding
  • Prioritized recommendations
  • Suggested actions (ask user before implementing)

Extended Review Capabilities

Beyond the basic procedure, this skill can:

Validation Levels

Level 1: Basic Compliance

  • Check if major features are implemented
  • Verify core functionality exists

Level 2: Detailed Verification

  • Validate all specified behaviors
  • Check edge cases and error handling
  • Verify data structures and interfaces

Level 3: Comprehensive Audit

  • Deep dive into implementation details
  • Compare algorithms and approaches
  • Validate performance requirements
  • Check all specified constraints

Cross-Reference Analysis

  • Bi-directional Check:

- Specs → Code (what's missing in code?) - Code → Specs (what's not documented?)

  • Dependency Validation:

- Check if dependencies match specs - Verify integration points

  • Version Alignment:

- Ensure code version matches spec version - Track changes across versions

Test Coverage Integration

  • Compare test cases with spec requirements
  • Identify untested specified behaviors
  • Suggest test scenarios based on specs
  • Validate test coverage completeness

Report Format

Reviews generate structured reports:

# Spec-to-Code Review Report

## Summary
- Specifications Reviewed: X
- Code Modules Reviewed: Y
- Consistency Rate: Z%

## Checklist Summary
- ✅ Fully Implemented: A items
- ⚠️ Partially Implemented: B items
- ❌ Not Implemented: C items
- 🔍 Unclear/Needs Investigation: D items

## Critical Inconsistencies
[Items where code contradicts specs]

## Missing Implementations
[Specified features not found in code]

## Extra Implementations
[Code features not in specs]

## Recommendations
[Prioritized actions - ASK USER before modifying code]

Usage Examples

Example 1: Basic Spec Review

Use platonic-code-review to review the implementation of
RFC-0001.md (user authentication) in the src/auth/ directory.
Generate a report showing what's implemented and what's missing.

Example 2: Comprehensive Module Review

Use platonic-code-review to conduct a comprehensive review
of the entire API implementation against all RFCs in ./specs/.
Focus on API endpoints, data structures, and error handling.

Example 3: Feature Completeness Check

Use platonic-code-review to verify that all features
specified in RFC-0005.md (payment processing) are fully
implemented in src/payments/. Generate a detailed checklist.

Example 4: Gap Analysis

Use platonic-code-review to perform a gap analysis between
./specs/ and ./src/, identifying specifications without
implementations and implementations without specifications.

Important Notes

Default: Report Only

⚠️ This skill generates reports by default and does NOT modify code

When inconsistencies are found:

  1. Document the issue in the report
  2. Provide specific code references
  3. Explain the discrepancy
  4. Recommend actions
  5. ASK USER if they want code modifications

When to Modify Code

Only modify code when user explicitly requests:

  • "Fix the inconsistencies you found"
  • "Update the code to match the specs"
  • "Implement the missing features"

Otherwise, default to report generation only.

Integration with Other Skills

Works well with:

  • platonic-specs: Review code against RFCs managed by this skill
  • Documentation skills: Generate docs from review findings
  • Testing skills: Create tests for missing coverage

Dependencies

  • Read access to specification documents (RFCs, requirements, design docs)
  • Read access to source code
  • Understanding of project structure
  • Write access only if user requests code modifications

Best Practices

  1. Start with specs: Always read and understand specs first
  2. Be thorough: Check all specified requirements systematically
  3. Be specific: Provide exact file names, line numbers, and code references
  4. Be objective: Report facts, not opinions
  5. Prioritize findings: Critical issues first, minor ones later
  6. Ask before modifying: Never change code without user consent

See references/REFERENCE.md for detailed review procedures and examples.

平台分布

Codex

36.68%
按下载量换算50

Claude

28.78%
按下载量换算39

Cursor

19.55%
按下载量换算27

Gemini CLI

10.83%
按下载量换算15

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills