
Doublecheck Verification Pipeline
Makes the agent extract claims from its output, find sources, and flag likely hallucinations.
- Repository
- github/awesome-copilot
- Stars
- 40,703
- Licence
- MIT
- Files
- 2 files
What it does
Once installed, the agent checks its own factual statements before presenting them. It pulls out each verifiable claim (facts, numbers, citations, names, dates), looks for supporting or contradicting sources through web search, and marks each claim as verified, plausible, disputed, or a possible fabrication. For legal, regulatory, or compliance content, or when something looks wrong, it produces a fuller structured report with source links so you can check independently.
Great for
- People who use AI for research or summaries and want source links to check
- Students or analysts who need to confirm figures and citations before relying on them
- Teams producing legal or compliance-related drafts that need an extra verification pass
Skip it if
- Final legal or medical decisions: the skill verifies claims but cannot certify them, and search results may be incomplete or wrong
- Creative writing or code generation, where the skill itself says verification does not apply
- Offline or privacy-sensitive work: the skill relies on web searches, so content and claims may be sent to search providers
Try asking
- “Doublecheck that the statistics in this market summary are accurate”
- “Run full verification on the case citations you just gave me”
- “Verify this paragraph of AI-generated text and show me the sources”
- “Turn on doublecheck mode for the rest of this research session”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 2 original files, the MIT licence and an attribution note.
doublecheck.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip doublecheck.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip doublecheck.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Say "full report" on any response when the stakes are high, rather than relying on the inline summary
- Treat VERIFIED as a pointer to a source to read yourself, not as proof
- Paste the exact text you want checked for the most precise claim extraction
What's inside
- assets/verification-report-template.md
- SKILL.md
Redistributed unmodified from github/awesome-copilot under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.