agentleFS
sxntixgosxntixgo/frugal-skills/skills/online-research/SKILL.md

online-research

Skill0 starsChanged 4 months agoMITView on GitHub

Researches a question on the web with source-quality discipline — plans queries, gathers and evaluates sources, triangulates each claim across independent sources, and answers with citations. Use when asked to research, investigate, fact-check, or find current/authoritative information online.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. renfei-design/Figma-AI-Bridge★ 16same name

    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 repository11 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    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 WebSearch/WebFetch when they're available.

    line 12
  2. Distinctive:

    Search, then open the promising sources — don't judge a page by its snippet.

    line 42
  3. Distinctive:

    No fabricated sources, quotes, or figures.

    line 73

≈ “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

Online Research

Answer the question from evidence, not memory. Corroborate every load-bearing claim across independent sources and cite it. Uncertainty stated plainly beats a confident guess.

Use WebSearch/WebFetch when they're available; if only one (or a different retrieval tool) is present, apply the same method with what you have. Never fabricate a source, quote, or URL.

Scope

This skill DOES:

  • Plan queries, gather sources, judge their credibility, triangulate claims, and synthesize a cited answer with a confidence level and known gaps.

This skill does NOT:

  • State claims from memory as if researched — if you didn't verify it, say so.
  • Rely on a single source for a load-bearing claim.
  • Invent citations, quotes, dates, or numbers.

Steps

Surface as TodoWrite items. Checkpoint running notes and the source list to ./.claude-skills-work/online-research/<topic>/notes.md as you go, so a long research task can pause and resume from where you left off.

1. Clarify the question ⛔ STOP if ambiguous

State what's actually being asked, the scope, and what a complete answer needs. Note if it's time-sensitive (needs current info). If the ask is ambiguous, ask before burning searches.

2. Plan queries

Decompose into sub-questions. For each, list search angles and terms (include alternates/synonyms and, for current facts, a recency qualifier). Write the plan to the notes file.

3. Gather

Search, then open the promising sources — don't judge a page by its snippet. For each source, record in the notes file: URL, who published it, date, and the specific claims it supports. Prefer primary sources and originals over aggregators.

4. Evaluate & triangulate

For each source judge credibility, recency, and bias (see references/source-evaluation.md). Then, for each load-bearing claim, corroborate it across at least two independent sources (not two sites copying one wire story). Mark claims that are contested, single-source, or uncertain.

5. Synthesize with citations

Answer the question. Every load-bearing claim carries a citation to a source you actually read. State a confidence level and call out gaps, disagreements, and anything you could not verify. Distinguish established fact from your inference.

Red flags (stop and correct)

If you notice…Do this
Answering from memoryVerify against a source, or label it unverified
One source for a key claimCorroborate with an independent second source
Citing a page from its snippetOpen and read it before relying on it
Two "sources" that are the same wire storyFind a genuinely independent one
A confident tone on shaky evidenceState the uncertainty and the gap
A high-ranked but low-quality/SEO pageJudge credibility, not search rank
Stale info on a time-sensitive questionSeek current sources; note the date
Any invented URL, quote, or numberNever — remove it; only cite what you read

Done when

  • Each load-bearing claim is cited to a source you read and corroborated (or explicitly flagged single-source/uncertain).
  • The answer states a confidence level and lists gaps/disagreements.
  • No fabricated sources, quotes, or figures.

References (load only when needed)

  • references/source-evaluation.md — credibility signals, primary vs secondary, recency, bias, independence, and citation practice.

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.