writing-plans
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to generate implementation plans that include shell commands for running tests (e.g.,
pytest) and managing version control (e.g.,git commit). These commands are intended for subsequent execution and are derived from the input specification. - [PROMPT_INJECTION]: The skill creates a potential surface for indirect prompt injection by processing external technical specifications or requirements to generate executable implementation plans.
- Ingestion points: Technical specifications or requirements files are consumed as the primary input for plan generation.
- Boundary markers: The skill uses structured Markdown templates, headers, and specific checkbox syntax to delimit tasks and steps within the plan.
- Capability inventory: Generated plans can include instructions for file creation, codebase modification, and the execution of shell commands.
- Sanitization: The skill incorporates a mandatory review phase using a subagent template (
plan-document-reviewer-prompt.md) to validate that the generated plan aligns with requirements and identifies incomplete or missing components.
Audit Metadata