ray-weekly
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides shell commands in
references/project-map.mdto be executed by the agent. These commands iterate through the user's~/projects/directory and executegit logto extract commit metadata and activity statistics. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from git commit messages to generate summaries.
- Ingestion points: Git commit messages read from repositories in Step 1 of the analysis process.
- Boundary markers: Absent. The skill does not provide instructions to the agent to treat commit messages as untrusted data or to ignore instructions embedded within them.
- Capability inventory: Local shell execution (git) and terminal output.
- Sanitization: None. The skill relies on the agent's summary capabilities without explicit validation or escaping of the log content.
Audit Metadata