Research
Research a technical or product question using high-trust primary sources when a decision depends on current documentation, external evidence, compatibility details, standards, or verified behavior.From its SKILL.md
npx -y skills add martonpaulo/skills --skill 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
- 21 days oldThe repository was created 21 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 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
2.2 KB, 358 tokens by cl100k_base, as published. Nobody here has run it
Research
Use this skill when external evidence or current documentation materially affects a decision. For questions answerable from the local codebase alone, inspect the code directly.
For a build-versus-reuse decision, gather and verify external evidence here while dont-reinvent-the-wheel owns the recommendation.
Use apple-docs for authoritative Apple development documentation, deep-docs for precise version-aware software documentation, and context7 when an indexed snippet of a library's current API settles the point. This skill retains broader external investigation, comparisons, current events, product questions, academic evidence, and questions spanning multiple source types.
Workflow
- Define the question, decision it informs, and any version, platform, or date constraints.
- Search primary sources first: official documentation, specifications, standards, source code, release notes, first-party APIs, and original research.
- Verify information that may have changed. Record source dates and applicable versions when they matter.
- Cross-check consequential claims against a second authoritative source or the referenced implementation when practical.
- Separate the result into:
- confirmed facts supported by sources;
- reasonable inferences, labeled as such;
- uncertainty, conflicts, or missing evidence.
- Return a concise answer with citations or precise source references.
Persistence
Return findings in the conversation by default. Persist them only when they are likely to be reused or the user requests a durable artifact. Follow an existing repository convention; otherwise use docs/research/<descriptive-topic>.md. Create the directory only when writing the document.
Completion
Research is complete when the decision-relevant claims are cited, current where necessary, and explicit about remaining uncertainty. Do not modify production code as part of research.
What ships with it: 2 files
2.0 KB alongside SKILL.md
agents/
- openai.yaml193 B
- THIRD_PARTY_NOTICES.md1.8 KB