optise-helix-bluf-writer

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill demonstrates safe operational practices, including structured data handling and explicit verification steps.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes user-provided buyer prompts and voice samples. Evidence:
  • Ingestion points: Buyer prompts and long voice samples entering the context in SKILL.md (Step 1).
  • Boundary markers: None identified for untrusted text inputs.
  • Capability inventory: Capability to perform network requests via web_fetch and web search (mentioned in Operating Principles).
  • Sanitization: Strong enforcement of output structure via 6 patterns and word count validation (40-60 words), which limits the impact of potential injections.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 02:41 PM
Security Audit — agent-trust-hub — optise-helix-bluf-writer