agents-mcp

Pass

Audited by Gen Agent Trust Hub on Oct 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes local utility scripts that use subprocesses to interact with the system environment.
  • scripts/mcp_health_check.sh executes the claude CLI with subcommands like list and get to verify the connection status of MCP servers.
  • scripts/test_mcp_health_check.py executes /bin/bash to run the health check script against various mocked CLI outputs for regression testing.
  • [EXTERNAL_DOWNLOADS]: The documentation provides instructions for installing and running official and well-known tools from the MCP ecosystem.
  • Mentions of npx -y are used to guide the user in running the official @modelcontextprotocol/inspector and various servers from the official registry.
  • sources.json contains a comprehensive list of official specification pages, registries, and well-known vendor repositories (e.g., Anthropic, OpenAI, Sentry, Grafana) for documentation and discovery.
  • [DYNAMIC_EXECUTION]: A utility script uses an embedded script execution pattern to process data.
  • scripts/mcp_health_check.sh pipes an embedded Python script (via heredoc) into the interpreter. This script is used for parsing CLI output and stripping ANSI escape codes, ensuring a clean and secure interpretation of server statuses.
  • [SAFE]: The skill is primarily a security-centric framework that enforces best practices.
  • It includes detailed "Security Guardrails" and checklists for SQL safety, PII protection, and least-privilege access.
  • The health check scripts include privacy protections, such as redacting raw CLI output to prevent the accidental logging of credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 17, 2026, 08:26 AM