crewai

Pass

Audited by Gen Agent Trust Hub on Apr 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the 'crewai' Python package, which is a standard library for building multi-agent systems.
  • [PROMPT_INJECTION]: The skill templates for agents and tasks use direct variable interpolation (e.g., '{topic}') which can be exploited via indirect prompt injection if processing untrusted data. Evidence: 1. Ingestion points: Variables are ingested in 'config/agents.yaml' and 'config/tasks.yaml'. 2. Boundary markers: No delimiters or ignore instructions are present in the examples. 3. Capability inventory: The 'researcher' agent uses 'SerperDevTool' and 'WebsiteSearchTool' which perform network operations. 4. Sanitization: The examples do not include sanitization or validation of the interpolated variables.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 11, 2026, 06:18 PM
Security Audit — agent-trust-hub — crewai