lean-review
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions in
SKILL.md(Step 3) explicitly direct the agent to "execute targeted probes" involving shell commands to verify code behavior and requirements. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it processes untrusted external data and has command execution capabilities.
- Ingestion points: The skill ingests the "exact diff/revision" and "acceptance" criteria as specified in the "Input and result" section of
SKILL.md. - Boundary markers: There are no defined delimiters or instructions provided to the agent to disregard instructions potentially embedded within the code diffs or requirements.
- Capability inventory: The skill maintains the ability to execute shell commands via "targeted probes" as described in Step 3 of
SKILL.md. - Sanitization: The instructions do not include steps for sanitizing, escaping, or validating the untrusted input data before it is processed by the agent.
- Ingestion points: The skill ingests the "exact diff/revision" and "acceptance" criteria as specified in the "Input and result" section of
Audit Metadata