skills/agentskillexchange/skills/Scan agent workflows for tools, MCP exposure, and adversarial risk with Agentic Radar/Gen Agent Trust Hub
Scan agent workflows for tools, MCP exposure, and adversarial risk with Agentic Radar
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs users to install the
agentic-radarpackage and its framework-specific extras (crewai, openai-agents) via pip. These packages originate from an external organization not included in the pre-defined trusted list.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes agent workflows, tool definitions, and MCP server configurations to generate security reports, creating a surface for indirect prompt injection.\n - Ingestion points: Agent workflows, tool definitions, and MCP server configurations processed by the scanner (SKILL.md).\n
- Boundary markers: None mentioned; the instructions do not specify the use of delimiters or ignore-embedded-instruction warnings for the scanned content.\n
- Capability inventory: The skill maps tools, maps MCP servers, and generates shareable security reports using the
agentic-radarCLI tool.\n - Sanitization: No sanitization or validation of the scanned workflow content is documented in the skill instructions.
Audit Metadata