Skills
Navigate
skills/jeremylongshore/tons-of-skills-marketplace/cohere-security-basics/Gen Agent Trust Hub

cohere-security-basics

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is instructional and focuses on enhancing the security posture of applications using the Cohere API. It explicitly warns against hardcoding secrets and provides safe alternatives.
  • [SAFE]: Recommended practices include using environment variables for secrets, implementing .gitignore for .env files, and using official secret management tools from well-known services like AWS, Google Cloud, and Vercel.
  • [SAFE]: Includes specific logic for PII (Personally Identifiable Information) scrubbing and log sanitization to prevent accidental data exposure.
  • [INDIRECT_PROMPT_INJECTION]: The skill demonstrates handling untrusted user data in the safeCohereChat code example.
  • Ingestion points: The userInput parameter in the safeCohereChat function (SKILL.md).
  • Boundary markers: The instructions recommend utilizing Cohere's built-in safetyMode ('CONTEXTUAL' or 'STRICT').
  • Capability inventory: The skill demonstrates interaction with the Cohere Chat API.
  • Sanitization: The skill provides a scrubPII implementation using regular expressions to redact sensitive patterns (emails, phone numbers, SSNs) before sending data to the LLM.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 04:15 PM
Security Audit — agent-trust-hub — cohere-security-basics