Context7 research
Deep library, dependency, framework, and API documentation research using Context7 MCP tools only. Use when you need up-to-date docs, version-specific API reference, migration guidance, or to verify code against official documentation. Triggers on requests like “look up docs”, “API reference”, “how do I use X”, “latest/current”, “deprecated”, “breaking changes”, “best practices”, or whenever correctness depends on primary documentation.From its SKILL.md
npx -y skills add BjornMelin/dev-skills --skill context7-researchAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 5 stars5 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.1 KB, 621 tokens by cl100k_base, as published. Nobody here has run it
Context7 Research
Use Context7 as the source of truth for library/framework documentation during implementation, reviews, and debugging.
Workflow (MUST FOLLOW)
- Extract:
libraryName: product/package/framework name (e.g.,next.js,zod,supabase)objective: what you need to do (e.g., “configure SSR auth”, “migrate v3->v4”, “use hook X”)version: only if the user specifies one or you can infer it from repo context (otherwise default to “latest”).
- If the user already provided a Context7 library ID in
/org/project(or/org/project/version) format, skip resolution. - Otherwise call
mcp__context7__resolve-library-idwith:libraryName: extracted namequery: the full user objective (not just a keyword)
- Select the best match using this rubric:
- Exact/closest name match
- Highest source reputation
- Highest snippet coverage (prefer more snippets when reputation is similar)
- Highest benchmark score
- Versions available that match the requested version (if any)
- Call
mcp__context7__query-docswith the selectedlibraryIdand a tight query (see Query Budget). - Synthesize an answer:
- Include the
libraryIdused (and version if applicable) - Quote or paraphrase only what’s needed; prefer code examples
- If docs are missing/ambiguous, label the gap as
UNVERIFIEDand ask for the smallest clarification needed.
- Include the
Query Budget (HARD LIMIT)
Use at most 3 mcp__context7__query-docs calls per user request. Plan the queries up-front:
- Query 1 (overview): “Explain the concept and where it lives in the docs; include minimal example.”
- Query 2 (API details): “Exact function/class/method signatures and usage examples for the user’s task.”
- Query 3 (edge cases): “Pitfalls, errors, version differences, and recommended patterns.”
If you can’t cover everything within 3 calls, ask the user to narrow scope (or pick the highest-impact subquestion).
Output (DEFAULT SHAPE)
- Start with the recommended approach (1–5 bullets).
- Include at least one doc-backed example/snippet when the question is about code.
- End with “Assumptions / Gaps” if anything is uncertain (
UNVERIFIED) or version-dependent.
Resources
- Playbook:
references/playbook.md - Troubleshooting:
references/troubleshooting.md - Design + scoring (maintainers):
references/design-spec.md - Report template:
assets/report-template.md - Generate a new report file:
python3 scripts/new_report.py --out ./context7-research.md --library-name "<name>" --question "<objective>"
What ships with it: 6 files
8.2 KB alongside SKILL.md, 1 of them executable
agents/
- openai.yaml293 B
assets/
- report-template.md415 B
references/
- design-spec.md2.8 KB
- playbook.md1.9 KB
- troubleshooting.md1.1 KB
scripts/
- new_report.pyruns1.7 KB
Gives 0 of the 12 instructions most docs writing skills give in 621 tokens
Counted across 1,951 of the 3,904 authors here whose files we hold, read 2026-09-06
- Use third-person for skill descriptionsin 54 of 1951, across 35 files
- Start descriptions with Use whenin 43 of 1951, across 29 files
- Run baseline scenarios before writing any skillin 40 of 1951, across 26 files
- Use active voicein 40 of 1951, across 36 files
- Map file responsibilities before defining tasksin 36 of 1951, across 29 files
- Use checkbox syntax for tracking stepsin 35 of 1951, across 27 files
- Ask one question at a timein 35 of 1951
- Offer execution options after saving the planin 33 of 1951, across 24 files
- Include complete code in every stepin 33 of 1951, across 27 files
- Design units with clear boundaries and interfacesin 31 of 1951, across 23 files
- Announce the skill usage at the startin 30 of 1951
- Verify agent compliance after adding the skillin 29 of 1951, across 17 files
Said here and by no other author read
- Use Context7 as the source of truth
- Extract library name and objective
- Include library ID and version in answer
- Label missing or ambiguous docs as UNVERIFIED
- Provide code examples for implementation questions
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.