implement
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious obfuscation, persistence mechanisms, or unauthorized data exfiltration patterns were found.
- [COMMAND_EXECUTION]: The skill executes shell commands for testing and linting (pnpm typecheck, pnpm lint, pnpm test) and git operations (git diff, git commit). These are standard tools for its stated purpose of implementing code.
- [PROMPT_INJECTION]: An indirect prompt injection surface exists where arbitrary task descriptions are used to generate and execute code.
- Ingestion points: User-provided task arguments are used in SKILL.md.
- Boundary markers: No explicit boundary markers or instruction-guarding delimiters are used to isolate the task from system instructions.
- Capability inventory: The skill has the capability to execute shell commands via pnpm and git as defined in SKILL.md.
- Sanitization: No sanitization of task input or generated code is performed prior to execution, relying instead on the local test suite.
Audit Metadata