verification-gate

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from project tickets, specifications, and pull requests without explicit sanitization or boundary markers. While this is standard for a verification tool, it creates an attack surface for indirect prompt injection.
  • Ingestion points: Processes content from work-in-progress claims (Step 1) and original tickets or spec criteria (Step 5) in SKILL.md.
  • Boundary markers: The instructions lack specific delimiters or "ignore embedded instructions" warnings when interpreting external text.
  • Capability inventory: The skill instructs the agent to execute shell commands for running tests (Step 2) and verification checks like curl and CLI binaries (Step 4).
  • Sanitization: There are no requirements for validation or filtering of the content being verified before use in command execution or prompt interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 03:34 PM
Security Audit — agent-trust-hub — verification-gate