writing-plans
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for taking a 'Task Description' and generating a detailed implementation plan. This creates a surface where malicious instructions within the task description could lead to the generation of unsafe code or commands.
- Ingestion points: The 'Task Description' input (referenced in the Mermaid diagram and workflow in SKILL.md) is the primary ingestion point for potentially untrusted data.
- Boundary markers: The skill uses structured headings for the plan steps, but it lacks specific safety markers or instructions to disregard malicious directives embedded in the input data.
- Capability inventory: Implementation plans created using this skill can include arbitrary code blocks and shell-based verification commands (e.g., 'curl', 'npx').
- Sanitization: There are no instructions for the agent to sanitize or verify the safety of the implementation logic derived from the task description.
Audit Metadata