critic

Fact-checking AI drafts against external evidence involves grounding claims using web search, memory, and local files rather than self-reflection. This skill verifies high-stakes content drafts by fetching tool-grounded evidence and evaluating structural coherence to return a structured JSON verdict.

Tool-grounded verification works by cross-referencing draft claims against external sources fetched via web search, local files, and knowledge graphs. It enforces a rigorous verification process before finalizing content, eliminating AI hallucinations and structural errors in high-stakes documents.

Yes, you can verify regulatory filings and strategic proposals for structural coherence. This skill evaluates content drafts for logical flow, argument consistency, and professional diction, ensuring documents requiring high accuracy and logical integrity are structurally sound.

The best way to eliminate AI hallucinations in partner outreach documents is enforcing a rigorous, tool-grounded verification process. By grounding claims against external sources like web search and local files rather than self-reflection, hallucination risks are removed before finalizing.

Yes, read-only access to file systems, web search, and knowledge graphs is required for automated quality assurance. This skill needs these external tools to perform objective verification and fetch grounded evidence for high-stakes content drafts.

You should not use self-reflection for fact-checking content drafts when finalizing high-stakes documents like regulatory filings or strategic proposals. Self-reflection lacks external grounding, whereas tool-grounded verification eliminates hallucination risks by cross-referencing claims against fetched evidence.