recruitee-automation

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill directs the agent to connect to an external Model Context Protocol (MCP) server at https://rube.app/mcp and references toolkit documentation at composio.dev. These resources provide the tool definitions and schemas necessary for the skill's automation functions.- [COMMAND_EXECUTION]: Instructions provide a workflow for executing tools via RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH. These tools allow the agent to perform operations on the Recruitee platform based on schemas discovered at runtime.- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it fetches dynamic execution plans and tool schemas from a remote endpoint (RUBE_SEARCH_TOOLS).
  • Ingestion points: Remote data retrieved from RUBE_SEARCH_TOOLS (SKILL.md).
  • Boundary markers: None provided to distinguish between vendor instructions and potential data-embedded commands.
  • Capability inventory: Subprocess-like execution via RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH (SKILL.md).
  • Sanitization: No explicit validation or sanitization of the fetched tool schemas is described.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 05:43 AM