writing-plans
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external "spec or requirements" documents and translate them into a granular implementation plan. This workflow creates an indirect prompt injection surface where malicious instructions hidden in the input specification could be automatically incorporated into the generated plan's code blocks or shell commands.
- Ingestion points: The skill primarily processes an untrusted specification document in SKILL.md to derive its tasks.
- Boundary markers: There are no explicit instructions to use delimiters or security-focused boundary markers when interpreting the input spec.
- Capability inventory: The skill possesses the capability to write files to the local filesystem (under
docs/superpowers/plans/) and provides a mechanism for dispatching subagents or executing plans that contain generated shell commands (e.g.,git,python,pytest). - Sanitization: The instructions do not include steps for sanitizing or filtering the content of the input specification before it is used to generate tasks.
Audit Metadata