requesting-code-review
Dispatch code review subagents with focused context to catch issues before they compound.
github-repo-search
帮助用户搜索和筛选 GitHub 开源项目,输出结构化推荐报告。当用户说"帮我找开源项目"、"搜一下GitHub上有什么"、"找找XX方向的仓库"、"开源项目推荐"、"github搜索"、"/github-search"时触发。
Full skill instructions
从用户自然语言需求出发,经过需求挖掘、检索词拆解、GitHub 检索、过滤分类、深度解读,最终产出结构化推荐结果。
目标不是"给很多链接",而是"给用户可理解、可比较、可决策、可直接行动的候选仓库列表"。
stars >= 100、archived=false、is:public。60 次/小时。10 次/分钟(独立于 Core 额度)。硬性门控:环节一是整个流程的前置条件。无论用户的需求描述多么清晰,都必须走完本环节并获得用户明确确认后,才能进入环节二。禁止根据用户的初始描述直接推断需求并开始检索。即使用户说"直接搜就行",也要先输出需求摘要让用户确认。
目标:把"我想看看 XX"转成可执行、可排序、可解释的检索目标。
需确认信息(最少):
相关性优先 / 星标优先(默认:相关性优先)可直接使用的产品 / 可二次开发的框架 / 资料清单/方法论建议补充信息(可选):
阶段输出(固定格式):
核心诉求:
- 主题:xxx
- 数量:Top N
- 最低 stars:>= 100
- 排序模式:相关性优先 / 星标优先(默认:相关性优先)
- 目标形态:xxx
- 偏好:xxx(可空)
- 排除:xxx(可空)
向用户确认以上信息。用户明确确认后才能进入环节二,否则停在这里继续对齐。
目标:平衡"召回率"和"相关性",避免只靠单词硬搜导致偏题。
拆词规则:
每组 query 由以下维度组合:
产出格式:
Query-1: "xxx"
目的:高召回核心主题
Query-2: "xxx"
目的:补同义词盲区
执行原则:
owner/repo 去重。候选池字段(最少):
owner/repostarsdescriptionrepo_urlarchivedlanguageupdated_attopicslicense硬过滤(默认):
stars >= 100archived = falseis:public可选硬过滤(按需):
fork = falselanguage:xxx目标:解决"命中 memory 但其实不是 agent memory"的噪音问题。
噪音剔除规则(示例):
排序原则(V1.1):
star 不再作为主排序,只作为召回门槛之一。
建议综合排序权重:
目标:让用户一眼看懂"这个仓库到底是什么角色",避免把框架、应用、目录混为一谈。
推荐类型字典:
目标:不是"仓库简介复述",而是输出"对用户有决策价值"的详细介绍。
深读最低要求:
每个入选仓库至少查看:
项目介绍写作要求(固定):
"项目介绍"必须包含两部分并写细:
可补充:
交付结构(固定):
Top N 表格字段(固定):
| 仓库 | 星标 | 仓库归属类型 | 项目介绍(是什么 + 推荐理由) | 其它信息补充 | 链接 |
|---|
"其它信息补充"建议内容:
迭代触发条件:
用户反馈"太泛/太窄/不够准/解释不够细"。
迭代动作:
100Top 10archived=falseDispatch code review subagents with focused context to catch issues before they compound.
Evaluate code review feedback with technical rigor before implementing, avoiding performative agreement and blind implementation.
Isolated git worktrees with smart directory selection and safety verification.
Structured workflow for completing development branches with test verification and merge/PR options.
review
Conducts multi-axis code review. Use before merging any change. Use when reviewing code written by yourself, another agent, or a human. Use when you need to assess code quality across multiple dimensions before it enters the main branch.
Standardized git commits using Conventional Commits specification with intelligent diff analysis and message generation.
Structures git workflow practices. Use when making any code change. Use when committing, branching, resolving conflicts, or when you need to organize work across multiple parallel streams.
Code review focused exclusively on over-engineering. Finds what to delete: reinvented standard library, unneeded dependencies, speculative abstractions, dead flexibility. One line per finding: location, what to cut, what replaces it. Complements correctness-focused review, this one only hunts complexity.
Generate Excalidraw diagrams from natural language descriptions in multiple formats.
Whole-repo audit for over-engineering. Like ponytail-review, but scans the entire codebase instead of a diff: a ranked list of what to delete, simplify, or replace with stdlib/native equivalents. One-shot report, does not apply fixes.
Systematic code review practices for constructive feedback, bug detection, and team knowledge sharing.
Opus.ai: Revolutionize Your Web Experience
Build a no-code AI app in minutes.
An IDE for code migration from legacy to modern frameworks through coding agents.
Automate CGI animation in live-action scenes
Branded artistic QR-code concepts
Launch a website in seconds with AI.
Streamline Your Coding Experience with AI Code Helper
Convert any screenshot or design to clean code.