返回 Skill 列表
extension
分类: 开发与工程无需 API Key

rpi-research

进行深入的、有证据支持的代码库研究,包括并行子调查,并生成带有文件/行引用的结构化备忘录。每当用户要求研究、分析、追踪、调试、审计或理解代码是如何连接的时候使用此方法,包括诸如“分析这个仓库”、“追踪这个流程”、“这是在哪里处理的”、根本原因调查、架构深度探讨、行为比较和实现映射等请求,即使用户从未明确说‘研究’这个词。

person作者: jakexiaohubgithub

RPI Research

Use this skill to answer complex codebase questions by combining targeted file reading, parallel investigation, and synthesis into a saved research note.

Initial Response

When the skill is invoked, respond with:

I'm ready to research the codebase. Please provide your research question or area of interest, and I'll analyze it thoroughly by exploring relevant components and connections.

Then wait for the user's research query.

Workflow After Receiving the Query

  1. Read directly referenced files first.
  • If the user names specific files, read those files fully before delegating work.
  • Read in the main context first so sub-investigations start from accurate context.
  1. Decompose the question into research tracks.
  • Break the request into clear research areas (components, flows, patterns, ownership boundaries).
  • Track subtasks in a task list so coverage is visible and no branch is dropped.
  • Map likely directories/files before parallel execution.
  1. Run parallel sub-investigations.
  • Spawn multiple focused sub-agents/tasks to research different tracks concurrently.
  • Keep each sub-task narrow and concrete (for example: API entrypoints, data model usage, or error handling path).
  • If sub-agents are unavailable, execute the same tracks sequentially and keep notes separated by track.
  1. Synthesize findings after all tracks finish.
  • Wait for all active tracks to complete before writing conclusions.
  • Reconcile overlaps and conflicts across tracks.
  • Capture concrete file references with 1-based line numbers.
  • Highlight cross-component relationships and architectural decisions.
  1. Produce the research memo in this structure.
---
date: [Current date and time in ISO format]
researcher: [User's Name] / [AI Agent Name]
topic: "[User's Question/Topic]"
tags: [research, codebase, relevant-component-names]
status: complete
---

# Research: [User's Question/Topic]

## Research Question
[Original user query]

## Summary
[Direct answer with high-signal findings]

## Detailed Findings

### [Component/Area 1]
- Finding with citation (`path/to/file.ext:line`)
- Why it matters to the question
- Connection to other areas

### [Component/Area 2]
- Finding with citation (`path/to/file.ext:line`)
- Why it matters to the question
- Connection to other areas

## Code References
- `path/to/file.ext:line` - What is relevant there
- `path/to/other.ext:line` - Why it supports the conclusion

## Architecture Insights
[Design patterns, conventions, and notable tradeoffs]

## Open Questions
[Remaining unknowns and how to resolve them]
  1. Save and present the result.
  • Save research notes to rpi/research/TIMESTAMP_research_topic.md.
    • For TIMESTAMP, Use a Windows-safe timestamp for TIMESTAMP: YYYY-MM-DD-HH-MM (e.g., 2026-03-05-22-10).
  • Use a short snake_case topic slug for the filename suffix topic.
  • Reply to the user with a concise summary plus key file references.

Quality Bar

  • Prefer evidence over speculation; when inferring, label it clearly.
  • Cite enough references that another engineer can quickly verify conclusions.
  • Keep the memo self-contained so it is useful without reading raw task logs.
  • Prioritize high-signal findings over exhaustive but low-value dumps.