entelligence-automation

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill contains no detected high-risk patterns such as hardcoded credentials, malicious persistence, or unauthorized file access. The instructions are consistent with the primary purpose of automating Entelligence via Composio.
  • [EXTERNAL_DOWNLOADS]: The skill references the external MCP server endpoint https://rube.app/mcp as part of the setup instructions. This is a reference to the official Composio service infrastructure used to provide the toolkit capabilities.
  • [PROMPT_INJECTION]: The skill architecture involves a surface for indirect prompt injection as it dynamically ingests data from an external source.
  • Ingestion points: Dynamic tool schemas and recommended execution plans retrieved from the remote MCP server via RUBE_SEARCH_TOOLS.
  • Boundary markers: No explicit delimiters are specified to separate external content from internal system logic.
  • Capability inventory: Includes tool execution (RUBE_MULTI_EXECUTE_TOOL) and bulk operations (RUBE_REMOTE_WORKBENCH).
  • Sanitization: No validation or sanitization routines are provided for content fetched from the external service provider.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 03:37 AM