pre-handoff-verification

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted handoff content and artifact evidence, creating an attack surface for indirect prompt injection.
  • Ingestion points: Untrusted data enters the agent context via the candidate_delivery_package and proposed_truth_state_claims inputs in contract.json.
  • Boundary markers: The skill enforces a rigid 7-point verification rubric in SKILL.md and references/seven-point-checklist.md to gate final response generation.
  • Capability inventory: The skill itself is non-mutating and performs no command execution; it gates transitions to other skills like plain-language-response based on audit results.
  • Sanitization: The skill mitigates risks by implementing mandatory downgrade rules in references/downgrade-rules.md, requiring explicit validation notes, and forcing the recording of specific tool and command names to ensure provenance.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 03:04 PM