agentleFS
Sign inStart free
oleg494oleg494/coding-kit/skills/web-research/SKILL.md

web-research

Skill3 starsChanged 11 days agoMITView on GitHub

Use when a fact from outside the project is needed: "find out", "look up", "what do they say about", "how it works", "compare", "find information", "is this still true". Identify or read the primary source first — open a known URL directly, search only when the owner of the fact is unknown — then cross-check material claims, date them, and answer with links. Not for the local knowledge base or recorded project facts (use dev-wiki) or consequential tool/library decisions (use production-first-decisions).

45 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use data-analysis, meeting-notes and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • shanggqmSkill

    data-analysis

    数据分析技能，提供结构化的数据查询、统计分析和报告生成工作流。当用户需要查询数据库、进行数据统计分析、生成分析报告时使用。按规范输出原始数据、计算结果、Markdown报告和HTML可视化报告。

    Also in 5 projects that use this

  • claude-office-skillsSkill

    meeting-notes

    >

    Also in 4 projects that use this

  • MicrosoftSkill

    web-search

    Also in 5 projects that use this

  • kongyo2Skill

    prompt-engineering

    Comprehensive prompt engineering framework for designing, optimizing, and iterating LLM prompts. This skill should be used when users request prompt creation, optimization, or improvement for any LLM task, or when users need help translating vague requirements into effective prompts through collaborative dialogue and iterative refinement.

    Also in 4 projects that use this

  • rogue-dev-studioSkill

    browser-automation

    Canonical browser automation and web UI verification: Playwright, agent browser CLIs, and CDP-based tooling in one playbook.

    Also in 3 projects that use this

  • JetBrainsSkill

    brand-guidelines

    Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.

    Also in 4 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 38 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Identify or read the primary source: If you know who owns the fact (vendor docs, a repo, a spec, a law), open that URL directly with the …

    line 16
  2. Distinctive:

    Use enough evidence to resolve the question, not a fixed count of queries or pages.

    line 42

“Common” is measured: kept by at least 6 in 10 of the 38 edited versions of this file we could read. “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

Web Research — facts from the outside world

Workflow

  1. Define the missing external fact. Local code, project docs and recorded decisions do not need web corroboration; a fact already in the Wiki is retrieved from there.
  2. Identify or read the primary source. If you know who owns the fact (vendor docs, a repo, a spec, a law), open that URL directly with the reader. Search only when the owner is unknown, then navigate from the result to the original. Use a browser only for interactive or dynamically rendered content.
  3. Prefer primary over retelling. Official documentation, original research, source code, the standard text; check the actual version and whether the answer applies to it.
  4. Cross-check where material. One authoritative source settles a narrow fact; disputed, indirect or high-impact claims need independent evidence.
  5. Date the claim. State the version or date for anything time-sensitive; say what may have changed since.
  6. Answer with sources. Link every external claim and name what stayed unverified. Never fill a gap with invented facts: an unavailable source is a reported limitation, not a licence to guess.

When access fails (403/429/captcha/paywall)

Try another legitimate source or access path — the project's own mirror, the repo, a cached or archived copy, a different official page — without bypassing access controls, signing up, or paying automatically. If the fact stays unreachable, name it, list what you tried, and finish the parts the available evidence supports.

Research depth

Use enough evidence to resolve the question, not a fixed count of queries or pages. A definition can need one source; a comparison needs the alternatives, their versions and any benchmark that speaks to the decision. Stop when more search cannot change the answer.

Where the fact usually lives

Kind of factOwner to start from
API, library, CLI behaviorOfficial docs, README, changelog, source code
Industry practice, adoptionEngineering blogs, ADRs, issue threads, tech radars
Bug, error messageProject issue tracker, release notes, Q&A with an accepted answer
Tool or approach comparisonReproducible benchmarks, practitioner reports with method
Regulation, standard, licenceThe issuing body's own text (.gov, ISO/IEC, licence file)

Answer

  • Result first line. What was found, in one sentence.
  • Details with links. Each fact carries its source.
  • Caveats. Unverified parts, version sensitivity, conflicting sources.
  • Memory. Durable findings go through dev-wiki routing; a research run does not create its own memory policy.

Gotchas

  • The first search hit is often a retelling. Check who authored it and where the numbers come from.
  • SEO farms and AI-generated posts reproduce docs with errors. Go to the original.
  • Publication date is not the version date: a 2024 article may describe a behavior removed in 2026.
  • An unanswered issue thread is evidence of a problem, not of its fix; read the resolution state.

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.

Check your setup, free