Evidence Grading

Grade every factual claim by source class (A–E) before it is allowed to support a decision. Use whenever a plan, finding, opportunity, lead list, market claim, or recommendation depends on a fact pulled from the web, a document, or memory. Also use before calling surface_finding — nothing goes out ungraded. Triggers on "is this true", "how confident are we", "source", "verify", "the data says", "studies show", "I read that", and any number quoted without an N.

cspoerer62 Updated 0 repo stars

File contents

cspoerer62/agent-skills commit dbd3ba70c8

Frequently asked questions

npx get-skill add cspoerer62/evidence-grading