code-review
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill interpolates user-supplied arguments directly into the instructional context of subagents in
references/codebase-scan-workflow.mdandreferences/parallel-review-workflow.md. This represents an indirect prompt injection surface. - Ingestion points: The
$ARGUMENTSvariable is interpolated into the<tasks>and<scope>blocks within the reference files. - Boundary markers: The content is wrapped in XML-style tags (
<tasks>and<scope>), but there are no explicit instructions to the model to ignore or treat embedded commands as data rather than instructions. - Capability inventory: The skill possesses significant capabilities, including the ability to spawn subagents (
researcher,planner,code-reviewer) and execute shell commands viagit. - Sanitization: There is no evidence of input validation or escaping for the content provided in
$ARGUMENTSbefore it is placed into the prompt template.
Audit Metadata