claude-researcher

KO | EN

gitlite — search

#javascript #python #react #go #golang #machine-learning #deep-learning #macos #awesome #cli #kubernetes #vue #pytorch #ios

claude-researcher

(★ 13)

Structured research skill for Claude — parallel web search, test solutions, weighted scoring decisions

claude-researcher 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Claude Researcher Skill

Give Claude a research question. Get a structured report with tested solutions and a scored decision.

<a href="#install">Install in 30s</a> ·
 <a href="examples/vector-db-rag/decision/DECISION.md">See example output</a> ·
 ⭐ Star if useful

---

This skill gives Claude a structured research workflow: search the web in parallel, organize findings into folders, optionally test each solution, and produce a weighted scoring decision.

## What This Skill Does

- Creates a dedicated `<problem_name>_problem/` workspace.
- Searches for candidate solutions from multiple source types (community, articles, code repos, academic, reference docs, videos).
- Organizes each solution into its own directory with documentation notes.
- Produces a per-solution `README.md` with summary, sources, pros, and cons.
- Optionally tests each solution and records reproducible outcomes.
- Generates a final `decision/DECISION.md` with weighted scoring and a recommended choice.

## Install

```bash
mkdir -p ~/.claude/skills/researcher
curl -o ~/.claude/skills/researcher/SKILL.md \
  https://raw.githubusercontent.com/maher-naija-pro/claude-researcher/main/researcher/SKILL.md
```

Then restart your Claude CLI session so the skill is reloaded.

## How To Use

Ask for research tasks with prompts like:

- "Research solutions for fixing slow API response times."
- "Compare approaches for background jobs in Node.js."
- "Investigate options for vector databases for RAG."

The skill follows this flow:

1. Create problem directory.
2. Search for candidate solutions in parallel.
3. Create one directory per solution.
4. Gather source documentation into markdown files.
5. Write per-solution summary (`README.md`).
6. Optionally test each solution and write `test_results_<solution>.md`.
7. Create `decision/DECISION.md` with weighted scores and final recommendation.

## Workflow

```
┌─────────────────────────────────────────────────────────────────┐
│                     RESEARCHER SKILL WORKFLOW                   │
└─────────────────────────────────────────────────────────────────┘

Step 1 ─ Create Problem Directory
 mkdir <problem_name>_problem/
 → Confirm with user ✓

Step 2 ─ Search for Solutions in Parallel
           Subagents (Sonnet) search across:
           Community · Articles · Code · Academic · Reference · Video
           Present list of solutions found.
           → Confirm with user (may exclude solutions) ✓

Step 3 ─ Create One Directory per Solution
 solution_<name>/
 README.md
 documentation/

Step 4 ─ Gather Documentation per Solution
 Parallel subagents (Sonnet) — one per source category
 Each produces documentation/<doc_topic>.md

Step 5 ─ Document Each Solution
 Write solution README.md with:
 Summary · Sources · Pros · Cons
 → Ask: "Test each solution or compare docs only?" ✓
 │
 ┌──────────┴──────────┐
 │ Test │ Skip
 ▼ ▼
 Step 6 ─ Test Each Solution │
 Sequential subagents │
 (Sonnet, one at a time) │
 Write test_results_ │
 <solution>.md │
 └──────────┬──────────┘
 ▼
 Step 7 ─ Decision Summary (Opus)
 decision/DECISION.md with:
 Scoring table · Weighted aspects · Final recommendation
 → Present and confirm with user ✓
```

| Step | Action | Model | Confirm? |
|------|--------|-------|----------|
| 1 | `mkdir <problem>_problem/` | — | Yes |
| 2 | Parallel subagents — search (docs, GitHub, Reddit, SO, etc.) | Sonnet | Yes |
| 3 | `mkdir solution_<name>/` + `documentation/` per solution | — | — |
| 4 | Parallel subagents — gather docs into `documentation/<topic>.md` | Sonnet | — |
| 5 | Write `README.md` with summary, sources, pros/cons | — | Yes |
| 6 | Sequential subagents — test each solution, write `test_results_<name>.md` | Sonnet | — |
| 7 | Create `decision/DECISION.md` with scoring table and final decision | Opus | Yes |

## Output Structure

```text
<problem_name>_problem/
 solution_<name_a>/
 README.md
 documentation/
 <doc_topic>.md
 test_results_<name_a>.md
 solution_<name_b>/
 README.md
 documentation/
 <doc_topic>.md
 test_results_<name_b>.md
 decision/
 DECISION.md
```

## Example Output

See [`examples/vector-db-rag/`](examples/vector-db-rag/) for a complete run comparing vector database options for RAG, including per-solution docs, test results, and a final `DECISION.md` with weighted scoring.

## License

This project is licensed under the [MIT License](LICENSE).

claude-researcher

claude-researcher 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

claude-researcher

claude-researcher 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README