banana-illustration
Pass
Audited by Gen Agent Trust Hub on Jul 19, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) by interpolating untrusted user data into prompts. Ingestion points: $ARGUMENTS and user-provided descriptions. Boundary markers: Absent; no delimiters or ignore instructions are used to isolate user content. Capability inventory: Access to Bash (restricted to uv and ls) and the Read tool. Sanitization: Absent; the workflow does not include input validation or escaping steps for the user-supplied strings.\n- [COMMAND_EXECUTION]: The skill constructs shell commands by placing user-controlled strings directly into arguments for the
uv runcommand. This creates a vulnerability to command injection if the agent or the underlying shell execution environment does not strictly escape shell metacharacters.
Audit Metadata