Ensures rigorous factual accuracy through systematic, multi-pass evidence validation and source tiering.
The Iterative Verification skill implements a rigorous methodology to prevent AI hallucinations and superficial research. It enforces a structured loop that moves beyond single-pass searches by categorizing claims into specific evidence tiers—Verified, Credible, Alleged, or Speculative. By requiring strict thresholds such as multiple independent sources, evidence freshness, and deep-link tracing, it ensures that Claude only confirms information when it meets high-integrity criteria. This is particularly useful for deep investigative work, ethical investment analysis, and any scenario where factual reliability is mission-critical.
