code-review
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs standard
gitutilities (git diff,git log,git rev-parse) as defined inSKILL.mdto inspect the local repository's history and current changes.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection, as it incorporates untrusted data into its analytical workflow.\n - Ingestion points: In
SKILL.md, the agent is instructed to read data from git diff outputs, commit messages, and project documentation or specification files.\n - Boundary markers: Ingested content is passed to sub-agents without the use of explicit delimiters or instructions to ignore commands embedded within the code or specifications.\n
- Capability inventory: According to
SKILL.md, the skill uses theAgenttool to spawn sub-agents for analyzing the ingested data.\n - Sanitization: The instructions in
SKILL.mdlack any sanitization or validation logic to filter out potentially malicious instructions from analyzed files.
Audit Metadata