choplin/my-agent-skills/document-toolkit-fact-check
- community
Checking that the SKILL.md file is still in the repository.
Verifies the factual claims in a technical document against sources, and reports which hold, which do not, and which could not be checked.
Install
npx vskill@latest install choplin/my-agent-skills/document-toolkit-fact-checkSKILL.md
Fact Check Document
Ensure that technical documents contain ONLY verified facts - no assumptions, guesses, or unverified claims.
Execution — isolate the verification reads
Fact-checking reads many files, runs commands, and searches the web to confirm each claim. That bulk does not belong in the caller's context: if the agent can run a subagent, run the Process below in one and have it return only the findings. Otherwise apply the Process inline.
The procedure and the four-state output are the same either way — isolation is a context boundary, not a different check. Whichever way it runs, return the findings and leave rewriting the document to the caller.
Process
- Read the document thoroughly - Identify all factual claims, technical statements, and assertions
- Verify each claim by:
- Searching the codebase for implementation details
- Reading source files to confirm behavior
- Checking configuration files
- Running commands to verify system behavior
- Using web search for external libraries/APIs documentation
- Testing code snippets when applicable
- Mark findings as:
- VERIFIED: Confirmed as fact through code/documentation
- INCORRECT: Contradicts actual implementation
- UNVERIFIABLE: Cannot confirm (suggest removal or rewording)
- NEEDS UPDATE: Partially correct but needs clarification
- Report results with:
- Line-by-line analysis of claims
- Evidence from investigation (file paths, code snippets)
- Suggested corrections for any non-factual content
- Rewrite the document if requested, ensuring it contains only verified facts
Important Rules
- NO assumptions - If you cannot verify it, mark it as unverifiable
- NO opinions unless explicitly marked as such
- NO guesses about how things "might" or "should" work
- ONLY facts that can be proven through code, tests, or official documentation
- Every technical claim must have a verifiable source
When in doubt, mark it as unverifiable rather than making assumptions.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/document-toolkit/document-toolkit-fact-check/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/choplin/my-agent-skills/document-toolkit-fact-check)