refine-prose
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and refine untrusted user-supplied text such as emails, PR descriptions, and essays, which creates a surface for potential indirect prompt injection.
- Ingestion points: The skill accepts and processes user prose through multiple refinement passes defined in SKILL.md.
- Boundary markers: There are no explicit delimiters or instructions for the agent to ignore instructions embedded within the text being edited.
- Capability inventory: The skill instructions include reading local configuration files (~/.claude/skills/write-as-pete/VOICE.md and ~/.claude/skills/clean-copy/SKILL.md) to inform style, but the skill lacks network, file-write, or arbitrary command execution capabilities.
- Sanitization: No input validation or filtering is specified for the data before it is processed by the agent.
Audit Metadata