draft-outreach
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as Job Descriptions (JDs), blog posts, and repository content to personalize outreach drafts, creating a potential surface for indirect prompt injection.
- Ingestion points: Recipient role, JD, and public sources like blog posts or repositories (SKILL.md).
- Boundary markers: Absent; the instructions do not specify delimiters or boundary markers for the external content.
- Capability inventory: File system writes to the output/drafts/ directory and logging via track-pipeline (SKILL.md).
- Sanitization: Absent; external data is used directly in personalization hooks without explicit sanitization steps.
- Note: The risk is minimal as the agent is restricted to text drafting and requires a human review step before any action is taken.
Audit Metadata