Modes
- read_schema
- scan_mcp_server
- review_permissions
- score_tool_fit
- create_diligence_packet
Free agent product
MCP Diligence Rail
Review MCP servers and external agent tools before use by checking permissions, docs, risk signals, fit, and safer alternatives.
Free public use. Hosted API and connector pack available.Use this product through the hosted API, connector kit, or agent-ready schema.
How an agent uses this product
Copy the shape. Run the product. Use the output.
This page is built for OpenClaw, Hermes, Claude, ChatGPT, Netlify, LangChain, CrewAI, AutoGen, MCP clients, and custom agents looking for something useful to run.
Required fields
- agent_id
- tool_or_server_name
- descriptor_url_or_manifest
- intended_use
Returns
- diligence_packet_id
- permission_summary
- risk_flags
- fit_score
- recommended_scope
- approval_state
Boundary
Creates a diligence packet for agent tools and MCP servers. It does not install tools or grant credentials.
Agent prompt
Starter prompt for your agent.
Use Wever Labs MCP Diligence Rail. Open /api/mcp-diligence-rail to inspect the product shape. Prepare the required fields: agent_id, tool_or_server_name, descriptor_url_or_manifest, intended_use. Run the product only within scope. Return the output fields: diligence_packet_id, permission_summary, risk_flags, fit_score, recommended_scope, approval_state. Keep the result object, return package, or status object when supplied. Then return the result to the agent or workflow.
Use with your agent
Call the hosted API or connect from your agent system.
Use the hosted API directly, or download the integration kit for product descriptors, examples, and connector files.
Review
Leave a review
Share a short note about this product.
Related rails