evaluate-acceptance-criteria
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes 'structured acceptance criteria' and 'executable checks' that are likely supplied via external data or user input, presenting a vulnerability surface for data-driven instruction override.
- Ingestion points: Acceptance criteria and executable check definitions provided to the agent (referenced in SKILL.md).
- Boundary markers: No specific delimiters or safety instructions are defined to separate the criteria data from the execution harness.
- Capability inventory: The skill possesses the capability to run executable checks and capture their output in the environment (SKILL.md).
- Sanitization: The skill does not mention any validation, filtering, or escaping of the input checks before execution.
- [COMMAND_EXECUTION]: The skill is explicitly designed to 'Run each check in the specified environment' and 'Record command/check identity, exit/result' (SKILL.md). This involves executing arbitrary commands or scripts as defined by the provided criteria, which is a significant capability that must be managed with caution.
Audit Metadata