sup-writing-plans
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of natural language instructions and templates designed to guide the agent's planning process. It does not include any executable scripts, network operations, or sensitive data access.
- [SAFE]: The configuration 'disable-model-invocation: true' in the YAML frontmatter serves as a security control that limits the agent's ability to trigger other tools or models autonomously during the planning phase.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it transforms external input into commands. This is the intended functional purpose of the skill. 1. Ingestion points: Software specifications or requirements provided to the agent as documented in SKILL.md. 2. Boundary markers: None explicitly defined in the prompts to isolate input data from instructions. 3. Capability inventory: The generated plans involve writing Python code and executing CLI tools such as pytest and git as seen in the Task Structure section of SKILL.md. 4. Sanitization: No specific sanitization or filtering of the input specification is mentioned, although the secondary review process defined in plan-document-reviewer-prompt.md acts as a manual oversight mechanism.
Audit Metadata