khe-content-engine
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content, such as previous social media posts and newsletter archives, to sample a writing voice. This ingestion of external data theoretically allows for indirect prompt injection if the source material contains malicious instructions designed to override agent behavior. However, the skill implements significant mitigations: Hard Rule 1 strictly forbids the agent from publishing or scheduling posts autonomously, requiring a person to manually approve every draft. Additionally, Hard Rule 2 prevents the agent from inventing facts, numbers, or quotations, and a local validation script (scripts/content-check.mjs) is used to scan outputs for AI-writing "tells" and consistency with the provided brief.
Audit Metadata