plan-writing
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION] (SAFE): No override markers or jailbreak patterns were found; the skill is strictly focused on task organization principles.
- [DATA_EXFILTRATION] (SAFE): No hardcoded credentials or access to sensitive file paths (e.g., .aws, .ssh) detected.
curlis used only as an example for API testing. - [REMOTE_CODE_EXECUTION] (SAFE): No remote scripts are downloaded or executed. References to
.pyscripts are illustrative examples of project-specific tools. - [INDIRECT_PROMPT_INJECTION] (LOW): The skill uses user-provided task descriptions to generate file names in the project root. Ingestion points: Task names/slugs derived from user input. Boundary markers: Absent. Capability inventory: Writing markdown files to the local file system. Sanitization: Not specified in the instructions. This is a standard functional surface for a planning tool.
Audit Metadata