agentleFS
DeL-TaiseiOzakiDeL-TaiseiOzaki/claude-code-orchestra/.claude/skills/research-lib/SKILL.md

research-lib

Skill199 starsChanged 2 months agoMITView on GitHub

Research a library and create comprehensive documentation in .claude/docs/libraries/.

2 other projects use a version of this, 2 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use general-purpose-opus with WebSearch/WebFetch for comprehensive library research.

    line 59
  2. Distinctive:

    If the subagent is unavailable, verify via manual web search:

    line 79

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Research Library

Research $ARGUMENTS and create documentation in .claude/docs/libraries/.

Scope: this skill produces documentation only. It never edits source code and never edits a dependency manifest or lockfile — if the research implies a version bump, report it as an action item instead of making it.

1. Resolve the Document Path

Derive the path once, from the script, and reuse the value everywhere below. $ARGUMENTS is raw user wording (FastAPI, ruamel.yaml, @scope/pkg); the script normalizes it the same way lib_inventory.py normalizes a declared dependency, so the doc filename and the dependency name always match.

python3 .claude/skills/_shared/workspace.py --skill research-lib \
  --title "$ARGUMENTS" --create

Emits {ok, skill, slug, paths: {lib_doc}, dirs, created, artifacts, verify}. Exit 0 on success, 1 on a bad argument. Use paths.lib_doc verbatim from here on — never re-derive it by hand, and never substitute $ARGUMENTS.md.

2. Resolve the Version This Project Uses

The declared and locked versions are facts with one correct answer, so read them instead of asking the web:

python3 .claude/skills/update-lib-docs/lib_inventory.py --library "$ARGUMENTS"

The dependencies array carries {name, declared_spec, declared_in, locked_version, locked_in, ecosystem}. > **Version Checked**: records locked_version when a lockfile pins one, otherwise the version satisfying declared_spec. If dependencies is empty the library is not (yet) a dependency of this project — record the version you are documenting against and say so in ## Constraints & Notes.

Exit codes: 0 ok, 1 bad argument, 2 a manifest could not be parsed (manifest_errors), 3 a library doc could not be read (read_error). On a non-zero exit, stop and report — do not fill the version in from a web search.

Newer upstream releases are research findings, not the project's version: note them under ## Constraints & Notes as an upgrade consideration. The blockquote is the only place a version number lives.

Research Items

Primary Tool: General-Purpose Subagent (Opus)

Use general-purpose-opus with WebSearch/WebFetch for comprehensive library research. Pass the resolved path and version into the prompt:

Agent tool:
  subagent_type: "general-purpose-opus"
  prompt: |
    Research: {library}. Find official documentation, key features,
    constraints, best practices, known issues, and usage patterns.
    Use WebSearch and WebFetch to gather information.
    The version this project uses is {locked_version or declared_spec};
    document against that version, and note newer releases separately.
    Save results to {paths.lib_doc} using the template in
    .claude/skills/research-lib/SKILL.md.
    List every URL you consulted under ## References.
    Return concise summary.

Fallback: WebSearch/WebFetch

If the subagent is unavailable, verify via manual web search:

  • Official documentation
  • GitHub README
  • PyPI / npm page
  • Latest release notes

Record every source you used in ## References; a research note whose sources are only in the conversation cannot be audited later.

Content to Document

  1. Basic Information
    • Official name, license
    • Official URL
    • Installation command
  2. Core Features
    • Main features
    • Basic usage (code examples)
  3. Constraints & Notes
    • Known limitations
    • Conflicts with other libraries
    • Performance characteristics
    • Async/sync considerations
  4. Usage Patterns in This Project
    • Recommended usage
    • Patterns to avoid
  5. Troubleshooting
    • Common errors and solutions

Output Location

paths.lib_doc from step 1 — the only derivation of this path.

Documentation Template

# {Library Name}

> **Last Updated**: {YYYY-MM-DD}
> **Version Checked**: {declared or locked version in this project}

## Overview

- **License**: {license}
- **Official URL**: {url}
- **Installation**: `{install command}`

## Core Features

{Description of main features}

## Basic Usage

```python
{Code example}

Constraints & Notes

  • {Limitation 1}
  • {Limitation 2}

Recommended Patterns

Do

{Good pattern}

Don't

{Anti-pattern}

Troubleshooting

{Error message}

Cause: {cause} Solution: {solution}

References

  • Official Docs
  • GitHub

The `> **Last Updated**` / `> **Version Checked**` blockquote is mandatory, not
decoration: `lib_inventory.py` parses exactly those two lines, and a doc without
them is invisible to `update-lib-docs` — never stale, never undocumented, never
maintained.

## Validate the Document

Both checks are gates. Do not report completion until both exit 0.

```bash
python3 .claude/skills/_shared/validate_doc.py --contract lib-doc \
  --file <paths.lib_doc>
python3 .claude/skills/_shared/workspace.py --skill research-lib \
  --slug <slug> --verify

validate_doc.py exit codes:

  • 1 — the file does not exist or could not be read. The delegated subagent did not write it (or wrote it elsewhere); the payload has no sections_missing. Re-delegate with paths.lib_doc; do not hand-edit a different path.
  • 2 — the file exists but violates the contract. sections_missing lists absent ## sections (Overview, Core Features, Constraints & Notes, References) and metadata_missing lists absent metadata lines (Last Updated, Version Checked). Fill those in and re-run.

workspace.py --verify exit 2 with verify.empty containing lib_doc means the file exists but is a stub (under 20 non-whitespace characters) — a summary was returned without real content.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.