eu-ai-act-specialist
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified. The skill operates as a deterministic compliance assistant following the EU AI Act framework.\n- [SAFE]: All included scripts (ai_system_risk_classifier.py, conformity_assessment_planner.py, and ai_act_obligation_tracker.py) rely exclusively on the Python standard library (argparse, json, sys), which eliminates risks associated with untrusted third-party dependencies.\n- [SAFE]: The skill lacks any network-facing capabilities, file-system modification logic, or credential-handling code, ensuring data remains local and secure.\n- [SAFE]: Indirect Prompt Injection risk is negligible: \n
- Ingestion points: User-provided JSON data via script arguments.\n
- Boundary markers: Data is handled as structured JSON, not interpreted as natural language instructions for the agent.\n
- Capability inventory: Scripts are limited to text processing and console output with no execution (eval/exec) or exfiltration capabilities.\n
- Sanitization: Input is processed through standard JSON parsing which provides inherent structural validation.
Audit Metadata