researcher

kind	SKILL.md
type	skill
title	Repository Researcher
category	workflows
tags	["research","repository-analysis","architecture-review"]
name	researcher
description	Use when analyzing repositories, conducting deep research on codebases, performing architecture reviews, or exploring large projects from a git URL.
author	xiuxian-artisan-workshop
date	2026-04-26T09:30-07:00
saliency_base	5.5
decay_rate	0.05
metadata	{"version":"2.1.0","source":"https://github.com/tao3k/xiuxian-artisan-workshop/tree/main/assets/skills/researcher","routing_keywords":["research","analyze","analyze_repo","deep_research","code_analysis","repository_map","sharded_analysis","architecture_review","git","repo","repository","github","url"],"intents":["Research repository","Analyze codebase","Deep research","Architecture review","Analyze git repo or link"]}

Researcher Skill

Sharded deep research for large repositories using the Qianji runtime (xiuxian-qianji) and a suspend/resume approval loop.

Architecture

┌─────────────┐     ┌────────────────┐     ┌──────────────────┐
│   Setup     │ --> │ Architect Plan │ --> │ Await Approval   │
│ clone + map │     │ shard proposal │     │ suspend/resume   │
└─────────────┘     └────────────────┘     └──────────────────┘
                                                    │
                                                    ▼
                                          ┌──────────────────┐
                                          │ Deep Analysis    │
                                          │ approved shards  │
                                          └──────────────────┘

Command

git_repo_analyer

Core command to execute repository research via Qianji.

Parameters:

• repo_url (string, required): Git repository URL to analyze.
• request (string, optional): Research goal. Default: "Analyze the architecture".
• action (string, optional): "start" or "approve". Default: "start".
• session_id (string, required for approve): Session returned by start.
• approved_shards (string, required for approve): Approved plan JSON string.

Execution model:

1. action="start":
   • clones and maps repository,
   • asks architect to propose shard plan,
   • returns session_id, proposed_plan, and approval prompt.
2. action="approve":
   • resumes same session with approved shard JSON,
   • runs deep analysis for approved shards,
   • returns final analysis payload.

Output

The command returns structured JSON. Typical fields:

• success
• session_id
• message / proposed_plan (start phase)
• analysis_result / full_context (approve phase)

Implementation Notes

• Runtime backend is xiuxian-qianji (Rust).
• Workflow definition is workflows/repo_analyzer.toml.
• No Python local runtime is maintained under this skill.

Files

researcher/
├── SKILL.md
├── README.md
├── references/
│   └── run_research_graph.md
├── workflows/
│   └── repo_analyzer.toml