code-review
Pass
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external code via the
$ARGUMENTSplaceholder. - Ingestion points: User-provided code changes are interpolated into the prompt for the
@principalagent. - Boundary markers: The skill uses section headers but lacks strict delimiters or explicit instructions to the agent to disregard instructions embedded within the code changes.
- Capability inventory: The skill can execute
gitcommands and write to local files viaobsidian_append_content. - Sanitization: No sanitization is performed on the input data. However, this risk is inherent to the skill's primary function of reviewing arbitrary code.
- [COMMAND_EXECUTION]: The skill utilizes dynamic context injection to include the current date and git commit hash in the review output. These commands are executed at load time, are strictly functional for metadata purposes, and do not incorporate user-supplied arguments.
Audit Metadata