writing-plans
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided specifications for multi-step tasks to generate implementation plans. These plans are explicitly designed to be consumed and executed by other agent skills (e.g.,
superpowers:executing-plansandsuperpowers:subagent-driven-development), creating a multi-step chain vulnerability. - Ingestion points: The skill ingests "specs for a multi-step task" from the user at runtime (SKILL.md).
- Boundary markers: Absent. The skill does not instruct the agent to use delimiters or to ignore potential instructions embedded within the user-provided specifications.
- Capability inventory: The generated plans involve creating and modifying files, executing shell commands via
pytestandgit, and spawning subagents for task execution (SKILL.md). - Sanitization: Absent. There is no requirement for the agent to sanitize or validate the user-provided specs before interpolating them into the implementation plan.
Audit Metadata