writing-plans
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external project specifications and generated plans, which creates an indirect prompt injection surface where untrusted input could influence generated implementation steps or shell commands. Evidence: The reviewer prompt in 'plan-document-reviewer-prompt.md' explicitly ingests untrusted SPEC_FILE_PATH and PLAN_FILE_PATH data. Boundary markers such as 'ignore instructions' tags are absent. Capability inventory includes the generation of code snippets and executable shell commands (e.g., git, pytest). No sanitization or validation of the input specification content is performed before processing.
Audit Metadata