fact-check · collaborative-deep-research — Claude Code skill: verify claims with web and academic sources
The fact-check skill breaks a claim into parts, searches for evidence in Google, Semantic Scholar and arXiv, then assigns a verdict to each part with cited sources.
Apache-2.0 276 installs ★ 51 updated 2026-03-22
npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a claude-codeWhat fact-check does
The skill uses paper-search to query web pages, academic papers and text snippets, then reads sources with paper read and paper-search browse. Each sub-claim gets a verdict: supported, partially supported, unsupported or uncertain. The final report lists all URLs and notes source quality.
Without the paper-search CLI installed, the skill will not run. It requires Bash, Read, Glob, Grep and Write tools.
It prefers primary sources — original papers and official data — over secondary summaries.
What it helps with
- verify a library or framework performance claim
- find academic sources for a statement in documentation
- break down a disputed technical claim and gather evidence for each part
- distinguish between no evidence found and evidence that contradicts
Example prompts for the agent
- Check if React 19 added automatic component memoization
- Fact-check: PostgreSQL is faster than MongoDB for join queries
- Find sources for the claim that WebAssembly supports multithreading
How to install fact-check
1Skills CLI (Vercel)
npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a claude-code2Manually: put the files in place
~/.claude/skills/fact-check/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/fact-check/
1Skills CLI (Vercel)
npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a codex2Manually: put the files in place
.agents/skills/fact-check/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a cursor2Manually: put the files in place
.cursor/skills/fact-check/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install collaborative-deep-research/agent-papers-cli .claude/skills/fact-check2Manually: put the files in place
.github/skills/fact-check/SKILL.md1Skills CLI (Vercel)
npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a gemini-cli2Platform CLI
gemini skills install https://github.com/collaborative-deep-research/agent-papers-cli --path .claude/skills/fact-checkManually: put the files in place
.devin/skills/fact-check/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about fact-check
How to install fact-check in Claude Code?
Run npx skills add collaborative-deep-research/agent-papers-cli --skill fact-check -a claude-code, then check ~/.claude/skills/fact-check/SKILL.md.
What does the fact-check skill do?
It breaks a claim into verifiable parts, searches for evidence in Google and Semantic Scholar, reads the sources and returns a verdict with citations.
What does fact-check need to work?
The paper-search CLI installed and access to Bash, Read, Glob, Grep, Write tools. Without paper-search, the skill will not start.
Source text
SKILL.md · Apache-2.0
--- name: fact-check description: Verify a specific claim by searching for evidence across web and academic sources. Use when the user asks to verify, fact-check, or confirm a statement. argument-hint: [claim to verify] allowed-tools: Bash, Read, Glob, Grep, Write --- Fact-check the following claim: "$ARGUMENTS" ## 1. Decompose Break the claim into specific, verifiable sub-claims. List them explicitly before searching. ## 2. Search for Evidence For each sub-claim: ``` paper-search google web "<sub-claim as question>" paper-search semanticscholar snippets "<sub-claim keywords>" paper-search semanticscholar papers "<sub-claim keywords>" --limit 5 ``` ## 3. Verify Sources For each promising source: ``` paper read <arxiv_id> <relevant section> # for academic papers paper-search browse <url> # for web pages ``` Prefer primary sources (original papers, official data) over secondary reports. ## 4. Assess For each sub-claim, assign a verdict: - **Supported**: strong evidence from multiple reliable sources - **Partially supported**: some evidence, with caveats - **Unsupported**: no evidence found, or evidence contradicts the claim - **Uncertain**: insufficient evidence to judge ## 5. Report Present: - The original claim - Each sub-claim with its verdict and supporting evidence - An overall assessment - All sources cited with URLs ## Guidelines - Always cite specific sources — never state a verdict without evidence. - Distinguish between "no evidence found" and "evidence contradicts." - Note the quality and recency of sources. - If a claim is about a specific paper, read that paper directly rather than relying on summaries.
© collaborative-deep-research · Apache-2.0 · collaborative-deep-research/agent-papers-cli
Similar skills
All skills for Science and research →fact-check by other authors comes first.