context-engineering

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is educational in nature, providing templates and strategies for curating agent context to improve output quality.
  • [SAFE]: It promotes security best practices by explicitly instructing against committing sensitive files like .env or hardcoded secrets.
  • [SAFE]: It addresses indirect prompt injection risks by providing a hierarchy of trust for source files and advising that external or user-generated content be treated as data rather than directives.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 08:39 AM
Security Audit — agent-trust-hub — context-engineering