verification-before-completion
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to read and evaluate the output of verification commands such as tests, builders, and linters. This creates an attack surface where maliciously crafted command output could potentially influence agent behavior. However, this risk is inherent to agent tool use and the skill itself primarily promotes increased scrutiny.\n
- Ingestion points: Output from external verification commands processed during the 'Gate Function' workflow.\n
- Boundary markers: No specific delimiters or warnings to ignore instructions within command output are provided.\n
- Capability inventory: The skill assumes the agent can execute shell commands and read their output to verify work.\n
- Sanitization: No filtering or sanitization of command output is defined.\n- [SAFE]: The skill contains no executable scripts, network exfiltration, or hardcoded credentials. The instructions focus on establishing a meta-policy for agent honesty and rigorous verification of work status, which aligns with secure development practices.
Audit Metadata