content-editor
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing untrusted user data (raw notes and brainstorms) to generate structured output. This inherent task creates a potential surface for indirect prompt injection.
- Ingestion points: User-provided raw notes and brainstorms as described in the transformation pipeline (Step 1: Ingest).
- Boundary markers: The skill includes a specific safety protocol requiring the agent to 'Wait for confirmation before transforming,' ensuring a user reviews the proposed document type and tone before any file operations occur.
- Capability inventory: The skill is scoped with Read, Write, Edit, Glob, and Grep tools to manage file transformations.
- Sanitization: While the skill applies structural and stylistic filters (Zinsser/Orwell rules), it does not explicitly sanitize raw input for hidden instructions, relying instead on the confirmation workflow.
Audit Metadata