two-pass-review
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code artifacts and diffs as its primary input in both the review and verification stages. Maliciously crafted content within these artifacts, such as instructions hidden in code comments, could attempt to influence the autonomous behavior of the spawned
code-reviewerorverifieragents. - Ingestion points: Artifacts (file or diff) processed in
SKILL.md(Pass 1 and Pass 2). - Boundary markers: None explicitly specified in the skill instructions.
- Capability inventory: No executable code or tool calls are defined within the skill itself, though the agents it spawns may possess their own capabilities.
- Sanitization: None specified.
- [NO_CODE]: The skill consists entirely of markdown instructions and does not ship with any scripts, executables, or configurations that would execute on the host machine.
Audit Metadata