workflow-guide
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown documentation and guidance files. It does not contain any executable scripts, binaries, or active code components.
- [PROMPT_INJECTION]: The skill defines a workflow where the agent's behavior is directed by content in a project file (Plans.md), creating a surface for indirect prompt injection. 1. Ingestion points: The agent reads the Plans.md file to detect tasks marked with pm:依頼中 or cc:TODO. 2. Boundary markers: The task list uses standard Markdown formatting which lacks robust delimiters to isolate task instructions from system prompts. 3. Capability inventory: The agent is instructed to use the /work command to implement changes and run tests based on the task descriptions. 4. Sanitization: No sanitization or validation steps are described for the task content before the agent executes the implementation steps.
Audit Metadata