prompt-proximity-architecture
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides instructions for designing a prompt proximity architecture, defining categories and bands for testing AI visibility. No malicious patterns, obfuscation, or unauthorized access attempts were detected.
- [SAFE]: The skill uses local JSON files for input (icp_hypotheses.json, buyer_jobs.json) and produces a structured JSON output. There are no network operations or sensitive file accesses.
- [SAFE]: Instructions are focused on logical categorization and allocation of prompt testing budgets, with no evidence of prompt injection or persistence mechanisms.
Audit Metadata