quick
Pass
Audited by Gen Agent Trust Hub on Apr 26, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes source code from the local environment to understand requested changes. Malicious instructions embedded within the code (e.g., in comments or string literals) could potentially manipulate the agent's behavior during the task.\n
- Ingestion points: In Step 1, the agent reads relevant code and affected files to understand the context of the fix.\n
- Boundary markers: Absent; there are no instructions provided to the agent to treat code as untrusted data or to ignore embedded directives.\n
- Capability inventory: The skill can execute shell commands (pnpm, mkdir, touch) and perform git commits.\n
- Sanitization: Absent; the skill does not validate or sanitize the content of the files it reads before processing them.\n- [COMMAND_EXECUTION]: The skill executes standard development environment commands to verify changes and manage internal state. Specifically, it uses
pnpm typecheck,pnpm check, andpnpm fixfor code verification, and standard filesystem utilities likemkdirandtouchto create a local context marker.
Audit Metadata