catch-up
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the conversation transcript and git repository which could contain instructions meant to manipulate the agent's summary output. 1. Ingestion points: conversation history, git log, and git diff as specified in SKILL.md. 2. Boundary markers: The instructions do not provide explicit delimiters or warnings for the agent to ignore instructions found within the repository data. 3. Capability inventory: The skill is limited to read-only shell commands (git log, diff, status) and text generation, with no file-write or network capabilities. 4. Sanitization: Content from the repository and conversation is processed without sanitization or filtering.
- [COMMAND_EXECUTION]: The skill utilizes shell commands to gather context about the project's state. Evidence: SKILL.md instructs the agent to run git log, git diff, and git status to confirm what changes have landed in the repository.
Audit Metadata