dcc-mcp/dcc-mcp-core

pip install dcc-mcp-core
curl -fL https://raw.githubusercontent.com/dcc-mcp/dcc-mcp-core/main/scripts/install-cli.sh -o install-cli.sh && sh ./install-cli.sh

Usage

• •dcc-mcp-cli dcc-types; dcc-mcp-cli list
• •dcc-mcp-cli search --query 'create sphere' --dcc-type maya --limit 20
• •dcc-mcp-cli describe <tool-slug>; dcc-mcp-cli call <tool-slug> --json '{...}'
• •Expose skills from Python with create_skill_server('maya', McpHttpConfig())
• •Add a skill with SKILL.md + tools.yaml (agentskills.io frontmatter contract)

Key features

• ✓Gateway with bounded tools/list (search, describe, load_skill, call) plus REST twin (/v1/search, /v1/describe, /v1/call)
• ✓Skill delivery: SKILL.md packages, progressive discovery, hot reload, marketplace distribution
• ✓DCC UI Control: bounded Computer Use-style automation with exact-window snapshots, semantic UIA + raw input fallback, Esc stop, audit trail
• ✓Main-thread affinity, multi-instance routing, async jobs, cancellation, checkpoints, workflows
• ✓Admin UI with calls, traces, logs, health, statistics
• ✓Python package (PyO3, Python 3.7-3.14) and Rust workspace

Best for

VFX/animation studios and TDs building reusable agent control planes for DCC and pipeline tools

Caveats

• ⚠Does not build agents itself; provides the infrastructure for agents to control DCCs
• ⚠Cannot safely preempt arbitrary code on a DCC main thread or guarantee rollback for non-transactional host APIs
• ⚠UI Control is a bounded fallback, not the default integration path