loonghao

Foundation library for the DCC Model Context Protocol (MCP) ecosystem. Provides Rust-powered action management, skills system, IPC transport, MCP Streamable HTTP server (2025-03-26 spec, with 2025-06-18 and 2025-11-25 awareness), sandbox security, shared memory, screen capture, USD scene support, and telemetry for AI-assisted DCC workflows. Use when working with Maya, Blender, Houdini, 3ds Max, or any DCC MCP integration.

Control live DCC hosts (Maya, Blender, Houdini, Photoshop, 3ds Max, and custom studio tools) through the dcc-mcp-cli command line. For ClawHub, OpenClaw, Cursor, Claude, and shell-capable agent hosts: verify gateway health, inventory registered DCC instances, search tools, inspect schemas, and invoke tools without speaking MCP directly. If dcc-mcp-cli is missing, ask for consent, download it from GitHub Releases, and fall back to Python stdlib REST only if download fails.

Infrastructure skill - guide developers and agents through creating or modernizing a full DCC-MCP adapter for Nuke, Blender, 3ds Max, Unreal, ZBrush, Houdini, Maya, and custom studio tools. Use when building server, dispatcher, gateway, packaging, and runtime integration. Not for authoring individual SKILL.md tool packages - use dcc-mcp-skills-creator.

Infrastructure skill - create, validate, scaffold, and review DCC-MCP skills for the dcc-mcp-core ecosystem. Use when authoring SKILL.md, tools.yaml, scripts, groups, prompts, or skill taxonomy. Not for creating a full DCC-MCP adapter repository - use dcc-mcp-creator.

Infrastructure skill — DCC-agnostic Qt UI introspection: list top-level windows, find widgets by name/class, describe widget properties, walk the widget tree, and poll for availability. Works in any DCC with a Qt binding (Maya, Blender, Houdini, Unreal, etc.). All tools are read-only and lazy-loaded.

Infrastructure skill - application UI observation and scoped action tools for DCC-adjacent workflows. Use app_ui__snapshot, app_ui__find, app_ui__act, and app_ui__wait_for when a host UI state is not exposed through native DCC APIs. Prefer DCC-native skills first, then use app_ui as a policy-controlled UI fallback.

Infrastructure skill - DCC-agnostic media probing, transcoding, thumbnails, frame extraction, and image-sequence-to-MP4 conversion through vx-managed FFmpeg. Use when agents need to inspect or share render/playblast outputs. Not for arbitrary shell or vx execution - use typed media tools only.

Infrastructure skill — DCC-agnostic observability primitives: generate error reports, capture screenshots, query audit logs, inspect tool performance metrics, and monitor process health. Works in any DCC environment (Maya, Blender, Houdini, Unreal, etc.) or standalone Python. Call dcc_diagnostics__error_report first whenever a tool fails with a vague error message. Not for primary task execution — use a domain skill for actual DCC operations.

Authoring stage - classic Maya dynamics primitives: rigid bodies, force fields, and field connections. Use for small simulation setup steps before baking with maya-animation. Not for keyframe editing (maya-animation), mesh modeling (maya-mesh-ops), or viewport output (maya-render).

Scene stage — create nodes, connect / disconnect attributes, query construction history, and inspect DG / DAG topology. Use whenever you reason about how nodes drive each other or construction history. Not for attribute *value* edits (use maya-attributes), polygon topology cleanup (use maya-mesh-ops), or scene file lifecycle (use maya-scene).

Pipeline stage — render globals and viewport capture: configure render settings, query them, capture playblasts. Use for producing final or preview imagery. Not for modeling (maya-mesh-ops), animation editing (maya-animation), generic file import/export (maya-geometry), or render farm submission (maya-render-farm).

Authoring stage — character / prop rigging: joints, IK, skin clusters, deformers, blend shapes, control curves, skin weights, constraints, and optional rig framework detection. Use when constructing rigs. Not for keyframe animation (maya-animation), pose libraries (maya-pose-library), or final scene assembly (maya-scene-assembly).

Scene stage — scene file lifecycle and DAG navigation. Use for new / open / save scenes, hierarchy queries, selection management, and top-level organisation. Not for mesh editing or interchange (FBX/OBJ): use maya-mesh-ops or maya-geometry instead.

Bootstrap stage — escape hatch for Maya work that has no packaged skill yet. Agents should prefer search_skills / dcc_capability_manifest → load_skill → typed tools (inputSchema + annotations) from domain skills; use execute_python or execute_mel only when no skill matches, for bulk in-process loops, or for API introspection. Includes typed plug-in lifecycle tools and introspection tools so an agent can discover flags and method signatures without leaving the loop.

Authoring stage — keyframes, timeline, animation curves, constraint baking, and curve I/O. Use whenever you create or edit time-based motion. Not for rigging setup (maya-rigging), pose library (maya-pose-library), or render output (maya-render).

Scene stage — read, write, lock, and create attributes on any Maya node. Use whenever you need to query or set an attribute value with a typed, validated tool instead of dropping into execute_python. Not for connecting attributes (use maya-node-graph) or scripting full procedures (use maya-scripting).

Blender reusable light rigs, HDRI/world setup, light grouping, and render-view controls

Blender material presets, shader assignments, texture images, and color-management helpers

Blender texture, lighting, ambient-occlusion, and transfer-map baking helpers

Blender local asset metadata, project context, publish manifests, and package prep

Blender scene, mesh, material, animation, and export readiness validation

Blender add-on development diagnostics, reloads, UI metadata, and environment checks

Blender rigid body, cloth, collision, and simulation cache tools

Blender Geometry Nodes group, modifier, and exposed input tools

Pipeline skill — generic, adapter-owned project and shot/package automation: set/query the project root, tag and read portable asset metadata, validate a scene for missing files / bad output dirs / dirty state / cook errors, collect file dependencies into a manifest (no copying), and export a shot/package manifest (frame range, cameras, output nodes, caches, written files). No dependency on private production services.

Authoring skill — create and edit cameras and lights, frame the Scene Viewer, and report active camera/view state through typed tools. Pair with houdini-render for viewport capture and ROP render execution.

Pipeline skill — viewport capture/flipbook plus ROP render settings and render execution through typed, runtime-aware tools. Exports report written_files, elapsed time, and warnings without hanging the host. Pair with houdini-camera-light to set up cameras and lights first.

Authoring skill — lookdev and shader-network operations: inspect materials, libraries, shader nodes and assignments; get/set shader parameters with type-aware validation; inspect and rewire shader connections; reset assignments; and manage adapter-owned material presets. Pair with houdini-materials for material creation/assignment.

Pipeline skill — typed timeline, keyframe, channel, and cache-baking tools. Query/set time, FPS and ranges; set/get/delete/list keyframes; inspect channels and expressions; export/import channel JSON; bake channels and trigger bounded simulation/cache renders. The canonical timeline API.

Interchange skill — load-on-demand import/export of Houdini geometry and scenes through production formats (native bgeo/geo, OBJ, Alembic, FBX, USD) plus filesystem probing. Use these typed tools instead of hand-building ROP/LOP/SOP networks with raw scripts.

Pipeline skill - run file-based Houdini Python automation, manage hip files, set timeline ranges, and build small node chains from structured specs. Use when orchestrating repeatable Houdini tasks. Not for inspecting a scene only.

HDA skill - install, inspect, create, cook, and save Houdini Digital Assets with typed parameters. Use when loading .hda/.otl assets or executing an HDA node as part of automation. Not for generic node edits.

Token-efficient command execution patterns for AI agents using vx. Use when running builds, tests, linting, GitHub operations, or any command that produces verbose output. Teaches agents to filter output cross-platform using vx-managed tools (vx rg, vx jq) instead of platform-specific syntax (Select-String, grep, findstr). Includes token measurement, savings tracking, and deep recipes for cargo, gh, pytest, and more.

Best practices for using vx effectively. Use when following recommended patterns for tool management, project setup, and team workflows with vx.

Complete vx CLI command reference. Use when looking up specific vx command syntax, flags, output formats, or token-efficient forwarding. vx-native commands support --json, --toon, --compact, and --output-format; forwarded tools keep native output unless compact mode is explicitly requested.

Troubleshooting guide for vx issues. Use when encountering installation failures, version conflicts, PATH issues, or other vx problems.

Teaches AI agents how to use vx, the universal dev tool manager. Use when the project has vx.toml or .vx/, or when the user mentions vx, tool version management, Git/GitHub operations, or cross-platform setup. vx auto-manages Node.js, Python, Go, Rust, and 142 providers via Starlark DSL provider.star files. Also covers MCP integration patterns and GitHub Actions.

Teaches AI agents how to use vx, the universal dev tool manager. Use when the project has vx.toml or .vx/, or when the user mentions vx, tool version management, Git/GitHub operations, or cross-platform setup. vx auto-manages Node.js, Python, Go, Rust, and 138 tools via Starlark DSL providers. Also covers MCP integration patterns and GitHub Actions.

Best practices for using vx effectively. Use when following recommended patterns for tool management, project setup, and team workflows with vx.

Complete vx CLI command reference. Use when looking up specific vx command syntax, flags, or output formats. All commands support --json for structured output and --output-format toon for token-optimized output.

Domain skill - move nodes and set positions in the current Autodesk 3ds Max scene. Use when the user asks to place, move, offset, translate, or position objects. Not for creating new geometry.

Domain skill - capture visual evidence from the current Autodesk 3ds Max viewport. Use when the user asks for a screenshot, visual proof, or a README image of the current scene. Not for rendering final frames.

Domain skill - create basic 3ds Max primitive geometry on the main thread. Use when adding boxes, spheres, cylinders, or planes. Not for mesh editing, import/export, or material assignment.

Set up dcc-mcp-3dsmax for an agent or operator: install 3ds Max Python dependencies with 3dsmaxpy, generate MCP host configuration, guide the user through starting the runtime inside 3ds Max, and run a first live-tool smoke prompt.

Domain skill - run explicit 3ds Max scripting checks and inspect host-side developer surfaces. Use for debugging automation, resolving nodes, and inspecting runtime symbols when typed tools are not enough.

Domain skill - create, inspect, and adjust 3ds Max cameras, lights, and simple review lighting rigs.

Domain skill - manage 3ds Max display layers, node display state, and user-defined custom properties.

Domain skill - run generic 3ds Max scene and asset-readiness validation checks before export, render, or handoff.

An autonomous code-assistant agent that can review code, answer technical questions, generate tests, and suggest refactors — powered by AgentVerse skills and an empathetic developer soul. Exposes MCP tools and is A2A (Agent-to-Agent) compatible.

Publish, discover, and manage AI skills, agents, workflows, souls and prompts from the AgentVerse marketplace. Use when working with the agentverse CLI to search/publish artifacts, authenticate, or manage AI agent ecosystem components.

Runs a suite of smoke tests against any REST API and reports status, latency, and failure details.

Reviews a pull request or diff and produces structured feedback on correctness, security, performance, and style.

Manage multi-container Docker environments using docker compose. Start, stop, inspect, and health-check services defined in a compose file — ideal for local dev, CI service orchestration, and integration test setup.

Fetch GitHub Pull Request diffs, metadata, and CI status using the gh CLI. Summarise changes, check review status, and post structured review comments.

Transform, filter, and reshape JSON data using jq — the lightweight command-line JSON processor. Ideal for extracting fields from API responses, transforming CI artifact payloads, and scripting data pipelines.

A battle-tested chain-of-thought prompt template that guides LLMs through multi-step reasoning. Supports zero-shot CoT, few-shot CoT, and Tree-of-Thought variants. Reuses the OpenAI prompt engineering and Anthropic prompt library standards.

Convert an existing Qt/PySide/PyQt desktop project (QWebEngineView-based UI, QMainWindow browsers, Qt DCC tools) into an AuroraView project so it ships a lighter, Rust-powered WebView and gets full MCP automation for free. Use this skill whenever the user asks to "migrate", "convert", "port" a Qt/PySide/PyQt app to AuroraView, or when they want MCP-controllable Qt tooling.

Integrates ProofShot visual proof recording with the self-improvement workflow. Use when: (1) E2E tests fail or detect visual regressions, (2) Agent needs to verify UI changes it made, (3) Running self-iteration loops (build → test → fix → retest), (4) Preparing PR proof artifacts for review.

Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks.

Guide for writing idiomatic Rust code based on Apollo GraphQL's best practices handbook. Use this skill when: (1) writing new Rust code or functions, (2) reviewing or refactoring existing Rust code, (3) deciding between borrowing vs cloning or ownership patterns, (4) implementing error handling with Result types, (5) optimizing Rust code for performance, (6) writing tests or documentation for Rust projects.

This skill produces an actionable architecture diagnosis report for a codebase. It guides the AI to scan the repository, map subsystem boundaries, identify design flaws with concrete evidence (files/functions), assess risks, and propose a phased refactor roadmap plus future extension scenarios. Use this skill when the team needs a "diagnosis report" to decide whether/how to refactor.

通过 MCP 工具测试 AuroraView Python API，发现设计缺陷和代码味道。 迭代式测试：启动服务 → 调用工具 → 发现问题 → 修复 → 重新测试。

This skill helps test and debug the local development version of AuroraView MCP Server. It guides AI to configure mcp.json, start the local MCP server, run tests, and troubleshoot issues. Use this skill when developing or debugging the auroraview-mcp package.

This skill helps create RFC (Request for Comments) documents for proposing new features, architectural changes, or significant enhancements to the project. It provides templates, structure guidelines, and best practices for writing effective technical proposals. Use this skill when planning major changes that need team review and discussion.

Best practices for using vx effectively. Use when following recommended patterns for tool management, project setup, and team workflows with vx.

Teaches AI agents how to use vx, the universal dev tool manager. Use when the project has vx.toml or .vx/, or when the user mentions vx, tool version management, or cross-platform setup. vx auto-manages Node.js, Python, Go, Rust, and 105 tools via Starlark DSL providers. Also covers MCP integration patterns and GitHub Actions.

Parse and extract structured data from invoice documents

Keyframe animation tools for Autodesk Maya — set, query, and bake keyframes

Adobe Photoshop image operations — create new documents, export, resize canvas, flatten, merge layers

Adobe Photoshop layer operations — create, delete, reorder, duplicate, rename, set opacity/visibility, fill, blend modes

Adobe Photoshop text layer operations — create, edit, style text layers

Adobe Photoshop document management — open, save, create, and query documents and layers

Domain skill — read-only scene and tool introspection for ZBrush. Use when you need active tool path, subtool inventory, or selection index. Not for sculpting, export, or subtool mutation — use zbrush-subtool or zbrush-interchange.

Domain skill — select and inspect subtools on the active ZBrush tool. Use when you need to switch the active subtool or read visibility/lock flags. Not for mesh export or arbitrary Python — use zbrush-interchange or zbrush-scripting.

Domain skill — export meshes from ZBrush to interchange formats. Use when you need OBJ export from the active subtool. Not for in-scene sculpt edits — use zbrush-subtool first to pick the correct subtool.

Thin-harness skill — controlled Python execution and session diagnostics inside ZBrush 2026.1+. Use when no typed ZBrush skill covers the task or you need version/subtool counts. Not for routine scene edits — prefer zbrush-scene or domain skills first.

Best practices for using vx effectively. Use when following recommended patterns for tool management, project setup, and team workflows with vx.

Complete vx CLI command reference. Use when looking up specific vx command syntax, flags, or output formats. All commands support --json for structured output and --format toon for token-optimized output.

Teaches AI agents how to use vx, the universal dev tool manager. Use when the project has vx.toml or .vx/, or when the user mentions vx, tool version management, or cross-platform setup. vx auto-manages Node.js, Python, Go, Rust, and 105 tools via Starlark DSL providers. Also covers MCP integration patterns and GitHub Actions.