agent-standards
Pass
Audited by Gen Agent Trust Hub on Feb 24, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists of documentation and meta-instructions for agent behavior. No executable scripts or binary files are included.
- [PROMPT_INJECTION]: The skill defines an architecture for ingesting external data, creating a surface for indirect prompt injection. Evidence: (1) Ingestion points: Ingesting code, traces, and terminal output (documented in 'references/cognitive-architecture.md'). (2) Boundary markers: Present; recommends structured context packing using markdown artifacts ('references/context-engineering.md'). (3) Capability inventory: References tool use for search and specialized agent delegation. (4) Sanitization: Present; mandates a 'Memory Gate' protocol for secret scrubbing ('references/memory-systems.md').
- [NO_CODE]: No source code files (Python, JavaScript, etc.) were found in the skill package.
Audit Metadata