mixture-of-experts

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests user-provided topics for analysis and passes them to specialized subagents. This creates a surface for indirect prompt injection where malicious instructions in the audited data could attempt to influence subagent behavior.
  • Ingestion points: User input defined in the workflow for identifying audit topics in SKILL.md.
  • Boundary markers: The instructions do not define specific delimiters to separate user data from subagent prompts.
  • Capability inventory: The skill can spawn subagents and potentially modify files during its execution phase.
  • Sanitization: No input validation or instruction-filtering mechanisms are described.
  • [COMMAND_EXECUTION]: Includes a standard installation command via npx for the author's repository, which is consistent with the skill's purpose.
  • [EXTERNAL_DOWNLOADS]: Fetches guidance and references from well-known platforms including OpenAI, Cursor, and Zed. These references are documented neutrally as they originate from established service providers.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 10:35 AM
Security Audit — agent-trust-hub — mixture-of-experts