multi-agent-architect
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows secure coding standards by explicitly instructing against hardcoding API keys and recommending the use of environment variables.
- [SAFE]: It includes defensive programming patterns, such as step-count guards to prevent infinite loops and allowlist validation for LLM-based routing decisions.
- [SAFE]: Security notes specifically address the risks of tool execution (e.g., Python REPL) and advise using sandboxed environments.
- [SAFE]: Data management guidance includes scoping session memory with TTLs to prevent memory leaks and cross-session data exposure.
Audit Metadata