crewai-multi-agent
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No malicious instructions or bypass attempts were identified.
- [Data Exposure & Exfiltration] (SAFE): API key examples in the troubleshooting guide use standard documentation placeholders (e.g., sk-...). No actual credentials or sensitive file paths are accessed.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): Installation commands refer to the legitimate crewai library and its official tools. No suspicious remote code execution or piped shell commands were found.
- [Indirect Prompt Injection] (SAFE): The documentation describes tools for processing external data (SerperDevTool, PDFSearchTool). As documentation, this does not represent an exploitable vulnerability within the skill itself. 1. Ingestion points: references/troubleshooting.md (SerperDevTool, PDFSearchTool). 2. Boundary markers: Not demonstrated in documentation snippets. 3. Capability inventory: Tool usage within agents. 4. Sanitization: Not demonstrated.
Audit Metadata