ai-disk-doctor-integrations
A read-only MCP server that exposes AI Disk Doctor Core's scan, model inventory, and history tools via local stdio, with no mutation or shell capabilities.
README
AI Disk Doctor Integrations
Universal local Agent Skill and non-destructive diagnostic MCP server for AI Disk Doctor Core.
This repository is an integration and distribution layer. It does not implement a second scanner, cleaner, risk engine, recovery model, policy system, history engine, or explainability engine. The local Core remains the execution and policy source of truth.
Alpha Scope
The MCP server exposes four tools:
| Tool | Behavior | Model-facing inputs | Core side effect |
|---|---|---|---|
core_status |
Check Core availability, required command surface, and compatibility provenance | None | None |
scan_summary |
Run aidisk scan --json and return bounded Core evidence |
category? |
Current Core CLI may save .aidisk/reports/scan-*.json |
ai_model_inventory |
Run aidisk models inventory --json with Core defaults and return bounded assets |
tool? |
None intended |
latest_diff |
Run Core-owned aidisk diff --latest --json and return bounded changes |
None | None |
I0.1 exposes no clean, restore, quarantine, delete, arbitrary shell, arbitrary executable, arbitrary rules/policy path, arbitrary reports directory, arbitrary filesystem mutation tool, or explainability MCP tool.
Core state used by I0.1
- Tested Core baseline: v1.7.0 at
52f31509394d2165cba8908da00a1036ba90479d. - Latest merged Public Core reviewed during I0.1:
33d741130b9c2bdd386cb96a25e0f7c70dd1bce7, which merged M1Cexplainability-v1. - M1C is not consumed by I0.1. The current explainability contract is exposed through the Rust application boundary, while the Core CLI has no explainability CLI contract for this integration to call.
Install
Prerequisites
- Node.js 18 or newer.
- AI Disk Doctor Core v1.7.0 on
PATHasaidisk/aidisk.exe, or setAIDISK_EXE. - This integration is tested against Core revision
52f31509394d2165cba8908da00a1036ba90479d. Runtime compatibility is checked bycore_status; current Core binaries do not expose a runtime git revision, so the exact revision is not claimed from runtime detection.
Install from a checkout:
git clone https://github.com/quzhiii/ai-disk-doctor-integrations.git
cd ai-disk-doctor-integrations
npm install
npm start
The process uses MCP stdio transport and should be started by an MCP client. The npm dependency set is locked in package-lock.json.
OpenCode
OpenCode supports project-local MCP declarations in opencode.json or opencode.jsonc:
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"ai-disk-doctor": {
"type": "local",
"command": ["node", "C:/absolute/path/to/ai-disk-doctor-integrations/src/server.js"],
"cwd": "C:/absolute/path/to/your-workspace",
"environment": { "AIDISK_EXE": "aidisk" },
"enabled": true
}
}
}
Set cwd to the user workspace where Core snapshots should be stored, rather than the integration checkout. Then use opencode mcp list or ask the agent to check core_status. The checked-in example is under adapters/opencode/opencode.jsonc.
OpenCode Agent Skills are discovered from .opencode/skills, .claude/skills, or .agents/skills. Copy or symlink skills/ai-disk-doctor into one of those locations for a project or user installation.
Qwen Code
Qwen Code supports local stdio MCP servers and extensions in current official documentation. Equivalent .qwen/settings.json entry:
{
"mcpServers": {
"ai-disk-doctor": {
"command": "node",
"args": ["/absolute/path/to/ai-disk-doctor-integrations/src/server.js"],
"cwd": "/absolute/path/to/your-workspace",
"env": { "AIDISK_EXE": "aidisk" },
"trust": false
}
}
}
The repository root is a Qwen extension package through qwen-extension.json; see adapters/qwen/README.md. The locally installed Qwen CLI in this environment was too old to expose documented MCP/extension subcommands, so client smoke is not claimed.
CodeBuddy / WorkBuddy
Exact packaging is deferred until an official, locally testable CodeBuddy / WorkBuddy client contract is available; see adapters/codebuddy/README.md. No Expert persona or vendor-specific risk logic is included.
TRAE
TRAE MCP and Skill docs exist, but exact project packaging is deferred until a TRAE client is locally available for validation; see adapters/trae/README.md.
Safety Boundary
- MCP transport is local stdio only.
core_status.server.modeisnon-destructive-diagnostic: the server exposes no destructive action, butscan_summarycan cause Core-owned snapshot persistence.- The server starts only the configured
aidiskexecutable with fixed allowlisted argv. - Model-facing inputs are limited to
scan_summary.categoryandai_model_inventory.tool. - No model-facing rules, policy, root, reports directory, executable, shell, cleanup, quarantine, restore, or delete parameter exists.
scan_summaryis non-destructive butreadOnlyHint: falsebecause current Core CLI scan persists a Core-owned snapshot. It isdestructiveHint: falseand does not modify user/workspace files.- Output is bounded: scan findings, model assets, diff changes, Core stdout/stderr capture, error evidence, and MCP text content all have hard limits.
- Model inventory is metadata-only according to the Core contract; this integration does not read prompt, transcript, source, document, token, cookie, credential, or model binary contents.
Agents must never bypass these rules by deleting paths with shell commands. If a future mutation flow is approved, it must use a separate Desktop-mediated authorization surface.
Development
npm install
npm run check
npm test
npm audit --audit-level=moderate
git diff --check
cargo fmt --manifest-path spikes/rust-direct-core/Cargo.toml -- --check
cargo run --manifest-path spikes/rust-direct-core/Cargo.toml
For a real Core smoke, install/build the pinned Core revision 52f31509394d2165cba8908da00a1036ba90479d, set AIDISK_EXE to that binary, and run npm test. CI includes a pinned-Core AIDISK_EXE smoke in addition to protocol-only Node tests and the Rust direct-Core spike.
License
The integration source is dual-licensed under MIT or Apache-2.0. See LICENSE-MIT and LICENSE-APACHE. Core licensing remains governed by the public Core repository.
Status
I0.1 Alpha. This repository is not the commercial Desktop and contains no accounts, billing, entitlements, telemetry, cloud synchronization, or proprietary Desktop code.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.