One Skill. One CLI. Every connected DCC.
The default agent entry is the public dcc-mcp Skill plus dcc-mcp-cli. The Skill supplies workflow knowledge; the CLI owns gateway lifecycle, structured discovery, typed execution, diagnostics, and updates.
Give your agent the project map
Paste this into an agent to orient it before a DCC-MCP task:
Learn the DCC-MCP ecosystem before acting. Read https://dcc-mcp.github.io/llms.txt first and use https://dcc-mcp.github.io/llms-full.txt only when more detail is needed. Route my request to exactly one path: operate a live DCC with dcc-mcp, create or modernize an adapter with dcc-mcp-creator, or create a specialized workflow Skill with dcc-mcp-skills-creator. Use https://dcc-mcp.github.io/marketplace to discover installable capabilities, https://dcc-mcp.github.io/showcase for proven prompt patterns, and https://dcc-mcp.github.io/ecosystem for repository ownership. Prefer typed dcc-mcp-cli discovery and follow returned next_step values. Diagnose failures from request_id before retrying, redact evidence, and never install, publish, file an external bug, or change my machine without explicit permission. Start by returning the project route, the capability you will use, and the exact next safe step.Install the matching Skill
| Intent | Skill |
|---|---|
| Operate a live DCC, discover tools, or search extensions | @loonghao/dcc-mcp |
| Create or modernize a complete adapter/runtime | @loonghao/dcc-mcp-creator |
| Create or improve a DCC-specific Skill package | @loonghao/dcc-mcp-skills-creator |
OpenClaw:
openclaw skills install @loonghao/dcc-mcpClawHub-compatible agents:
npx --yes clawhub@0.23.1 install @loonghao/dcc-mcpStart a new agent turn after installation so the runtime can load the Skill.
Follow the typed workflow
# Confirm the gateway and connected hosts.
dcc-mcp-cli health
dcc-mcp-cli list
# Search narrowly, then follow the returned next_step.
dcc-mcp-cli search --query "create sphere" --dcc-type maya
# Execute the returned tool slug with validated arguments.
dcc-mcp-cli call <tool-slug> --json '{"radius": 2.0}'Search is the routing contract. Do not guess tool names or load every backend schema into context. Perform the returned targeted load or describe step only when requested.
Diagnose before retrying
Keep the request_id from a failed call. Use the CLI's built-in evidence path:
dcc-mcp-cli doctor
dcc-mcp-cli stats --status failureThen discover and call dcc_feedback__report through the same search workflow. Review the public-safe issue report before sharing it. Never publish raw evidence or create an external issue without user authorization.