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
- ilhae-ai/claude-code-orchestratoredited, first seen Oct 20, 2026
- gaebalai/claude-code-orchestratoredited, first seen Jan 31, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 2
- gaebalai/claude-code-orchestrator★ 1879% same
- ilhae-ai/claude-code-orchestrator★ 1879% same
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository17 other files it gives its agents.
- .agents/AGENTS.mdAGENTS.md
- AGENTS.mdAGENTS.md
- .codex/AGENTS.mdAGENTS.md
- catchupSkill
- checkpointingSkill
- codex-systemSkill
- context-loaderSkill
- design-trackerSkill
- featureSkill
- initSkill
- planSkill
- simplifySkill
- spikeSkill
- tddSkill
- team-executeSkill
- troubleshootSkill
- update-lib-docsSkill
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
- Very specific: line 59
Use general-purpose-opus with WebSearch/WebFetch for comprehensive library research.
- Distinctive: line 79
If the subagent is unavailable, verify via manual web search:
≈ “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.
Files
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" --createEmits {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
- Basic Information
- Official name, license
- Official URL
- Installation command
- Core Features
- Main features
- Basic usage (code examples)
- Constraints & Notes
- Known limitations
- Conflicts with other libraries
- Performance characteristics
- Async/sync considerations
- Usage Patterns in This Project
- Recommended usage
- Patterns to avoid
- 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> --verifyvalidate_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 withpaths.lib_doc; do not hand-edit a different path. - 2 — the file exists but violates the contract.
sections_missinglists absent##sections (Overview,Core Features,Constraints & Notes,References) andmetadata_missinglists 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.
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
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.

