anti-deception
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were identified. The skill's logic focuses on improving response integrity and pushback against deceptive prompts.
- [EXTERNAL_DOWNLOADS]: The skill utilizes a vendor-provided MCP server (
ejentum-mcp) and interacts with the vendor's domain (ejentum.com). This communication is required to retrieve the integrity scaffolds described in the functionality. - [CREDENTIALS_UNSAFE]: Authentication is handled using an environment variable (
EJENTUM_API_KEY), which is a recommended and secure practice for managing service tokens.
Audit Metadata