メニュー
High-performance GitHub research orchestrator. Features search result caching, multi-dimensional analysis, and sub-agent delegation to minimize token usage and prevent context contamination.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Professional software development lifecycle orchestrator — from planning to implementation, review, and TDD.
Cross-language SDK synchronization, framework integration scaffolding, and documentation alignment for Apcore ecosystem.
Deep Project Intelligence & Technical Due Diligence Skill — comprehensive analysis of architecture, codebase, and strategic positioning.
Analyze a collection of documents to build a knowledge map, identify themes, find gaps, duplicates, conflicts, and staleness, then produce an organized analysis with improvement recommendations. Designed for document ecosystems (cross-repo docs, research collections, mixed-format doc repos) where the challenge is understanding the landscape, not auditing against code.
Audit existing project documentation for quality, completeness, consistency, and code alignment. Scans docs against the actual codebase to find stale references, missing coverage, internal contradictions, and quality issues. Generates a structured findings report with severity levels and actionable fix recommendations. Works on any project with a docs/ directory.
Lightweight project decomposition skill. Analyzes project scope through a brief interview and determines whether the project should be treated as a single feature or split into multiple sub-features. For multi-split projects, generates a project manifest listing sub-features, dependencies, and execution order.
| name | git-researcher |
| description | High-performance GitHub research orchestrator. Features search result caching, multi-dimensional analysis, and sub-agent delegation to minimize token usage and prevent context contamination. |
You can trigger this skill by asking me to research GitHub projects. Here are some examples:
topic:ai, llm, agent, etc.stars:>100, forks:10..50.--lang zh (Chinese report), --lang en (English report, default).search_results.json to avoid redundant API calls.generalist or codebase_investigator) within its specific directory.gh search repos with user criteria.search_results.json in the current directory.gh search repos "..." --json ... > search_results.jsonsearch_results.json first before re-searching.search_results.json and present a numbered list to the user.For each project, spawn a sub-agent (using the generalist tool) with the following specific mission:
codebase_investigator inside that directory to understand the Technical, User, and Investor perspectives.git-research.md report based on references/report_guide.md.codebase_investigator via a sub-agent.search_results.json and existing directories before every major action.--lang zh is used, ensure the sub-agent generates the report in Chinese.