Handbook

Task `doc_risk_classification` v1

Classify documentation claims by publication risk (low / medium / high) for the Forge ecosystem. High risk covers external commitments, benchmarks, security/compliance statements, and capability claims beyond stated…

Updated

Task doc_risk_classification v1

Summary

Classify documentation claims by publication risk (low / medium / high) for the Forge ecosystem. High risk covers external commitments, benchmarks, security/compliance statements, and capability claims beyond stated maturity. The task never changes claim text or status.

Inputs

Field Type Required Notes
claims array yes Claim objects (at least claim_text), typically from doc_claim_extraction.
temperature number no Default 0.05.
timeout_sec integer no Transport timeout override.

Output

JSON object with classified_claims: each item has claim_text, risk_level (low | medium | high), risk_reason, requires_owner_review (defaults true unless low risk).

Invariants

  • risk_level outside the enum fails with SchemaFailure (no silent coercion — risk is a gate input).
  • Claim text passes through unchanged.

Failure modes

  • SchemaFailure — empty claims, missing text, or invalid risk_level.
  • ParseFailure — invalid assistant JSON (one repair attempt).
  • TransportFailure — model/provider errors.

Changelog

  • v1 — Initial doc-hydration risk classification task (Hydration v2, Phase 5).