tech-hub-ai
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements proactive security measures for AI/ML workflows. It mandates a 'Security Lead' collaboration (skill sa-01) whenever user or personal data is involved and enforces explicit approval requirements for sensitive operations like modifying guardrails or accessing production environments.
- [PROMPT_INJECTION]: The orchestrator processes external task context and propagates it to subagents. The potential for indirect prompt injection is addressed through a structured execution model that uses explicit boundary headers, mandatory implementation checklists, and a validation step that verifies subagent outputs against predefined quality gates.
- [COMMAND_EXECUTION]: The agent coordinates the creation and modification of code and scripts. It mitigates the risk of unauthorized execution by categorizing tasks into automation thresholds, ensuring that changes to existing code, dependency additions, and production deployments require explicit user intervention.
Audit Metadata