modelbound avatar

Modelbound Mcp

modelbound/modelbound-mcp-server

A local-first MCP server that stores and syncs agent skills across Cursor, Claude, Kiro, Windsurf, VS Code, and Amazon Q. Exposes tools for validation, linting (front-matter, token counts, broken links), format conversion between IDE conventions, and diff operations. All core functions run offline over stdio with no API key required. Add MODELBOUND_API_KEY to unlock cloud sync, marketplace installs, and team resource trees grouped by platform hierarchy. Useful if you switch between AI coding tools frequently and don't want to rewrite rules and context files for each one. Also works as a standalone CLI for batch operations on skill directories.

tools
19
auth
Secret needed
transport
STDIO
env var
1
last updated
May 23, 2026

Tools

19 tools

Public tool metadata for what this MCP can expose to an agent.

files.createCreate a new context file on ModelBound3 params

Create a new context file on ModelBound

Parameters* required
contentstring
File body as text or markdown
context_typestring
File category (e.g. system_prompt, rule, doc)
titlestring
Human-readable title for the file
files.getGet the full content of a specific context file by ID1 params

Get the full content of a specific context file by ID

Parameters* required
file_idstring
UUID of the context file to retrieve
files.listList context files in your ModelBound library2 params

List context files in your ModelBound library

Parameters* required
context_typestring
Optional category filter
limitnumber
Max number of files to return (default 50)
files.searchSearch context files by keyword1 params

Search context files by keyword

Parameters* required
querystring
Keyword or phrase to search for in file titles and content
search.allSearch across all ModelBound content (files, agents, packs, corpora)1 params

Search across all ModelBound content (files, agents, packs, corpora)

Parameters* required
querystring
Keyword or phrase to search across all content types
optimization.tokensRun AI-powered token compression on context files (30-60% reduction)1 params

Run AI-powered token compression on context files (30-60% reduction)

Parameters* required
file_idstring
UUID of the file to optimize
security.scanSecurity scan for prompt/context files (prompt injection, secrets, etc.)1 params

Security scan for prompt/context files (prompt injection, secrets, etc.)

Parameters* required
file_idstring
UUID of the file to scan
optimization.healthGet health scores, staleness, and size warnings for all context files

Get health scores, staleness, and size warnings for all context files

No parameter schema in public metadata yet.

optimization.contextRetrieve context files ranked by relevance within a token budget2 params

Retrieve context files ranked by relevance within a token budget

Parameters* required
querystring
Task description used to rank relevance
token_budgetnumber
Maximum total tokens to return (default 128000)
optimization.auditToolsReport the token footprint of all enabled MCP tool definitions

Report the token footprint of all enabled MCP tool definitions

No parameter schema in public metadata yet.

optimization.costEstimate cumulative conversation cost and recommended reset points2 params

Estimate cumulative conversation cost and recommended reset points

Parameters* required
modelstring
Model identifier (e.g. gpt-5, claude-sonnet-4)
turnsnumber
Number of conversation turns to project
skills.listList portable ModelBound Skills1 params

List portable ModelBound Skills

Parameters* required
limitnumber
Max number of skills to return
skills.getGet a full Skill bundle including SKILL.md and bundled files1 params

Get a full Skill bundle including SKILL.md and bundled files

Parameters* required
skill_idstring
UUID of the skill to retrieve
corpus.createCreate a new RAG corpus for knowledge base search2 params

Create a new RAG corpus for knowledge base search

Parameters* required
descriptionstring
What this corpus contains
namestring
Corpus display name
corpus.searchSearch a RAG corpus using hybrid keyword + vector retrieval3 params

Search a RAG corpus using hybrid keyword + vector retrieval

Parameters* required
corpus_idstring
UUID of the corpus to search
limitnumber
Max results to return (default 10)
querystring
Search query
corpus.askAsk a natural-language question grounded in a RAG corpus2 params

Ask a natural-language question grounded in a RAG corpus

Parameters* required
corpus_idstring
UUID of the corpus to query
questionstring
Natural-language question
threads.startStart a new adaptive conversation thread1 params

Start a new adaptive conversation thread

Parameters* required
titlestring
Optional thread title
threads.appendAppend a message to an existing thread3 params

Append a message to an existing thread

Parameters* required
contentstring
Message text
rolestring
Message role: user, assistant, or system
thread_idstring
UUID of the thread
threads.getWindowGet the current context window for a thread1 params

Get the current context window for a thread

Parameters* required
thread_idstring
UUID of the thread

Sponsors

Configuration

MODELBOUND_API_KEYsecret

ModelBound API key for optional cloud sync features. Get one at https://modelbound.co/settings/api-keys

Readme

modelbound-mcp

Local-first MCP server for agent skills. Validate, lint, diff, and convert agent skill files across Cursor, Claude, Codex, Kiro, Windsurf, VS Code, and Amazon Q — no account required. Optional cloud sync with ModelBound.

npm License: MIT ModelBound Skill Checks

Why ModelBound?

AI tools come and go. You might use Cursor today, switch to Claude Code tomorrow, and try Kiro next week — but your skills, rules, and context shouldn't be locked into any one of them. ModelBound gives you a single place to store and manage your agent skills, so you can move between tools freely without rebuilding your setup each time. Write a skill once, sync it everywhere, and get more value out of every AI subscription you're already paying for.

What it does


View the full README on GitHub

Details

Repository
github.com/modelbound/modelbound-mcp-server
Category
AI & LLM Tools
Package
modelbound-mcp
Transport
STDIO
Auth
Required
Registry
MCP registry, active

Related AI & LLM Tools MCP servers

View all
1flipcoin-fun avatarFlipcoinflipcoin-fun/flipcoin-mcpAgent-first prediction markets on Base. Trade, quote, create markets via MCP.0stars
2mukundakatta avatarRagmetric Mcpmukundakatta/mcp-stackRAG retrieval IR metrics: recall@k, hit@k, MRR, NDCG@k, evaluate_batch.0stars
3mukundakatta avatarStreamparsemukundakatta/streamparse-mcpParse partial / truncated / messy JSON for LLM tool calls and structured outputs.0stars
4newmannn91 avatarMemdrive Mcpnewmannn91/memdriveLocal-first MCP memory: Markdown notes and images with full-text search for Claude and Cursor.0stars
5smara-io avatarSmarasmara-io/mcp-serverPersistent cross-platform memory for AI agents with Ebbinghaus decay scoring.0stars
6perceptdot avatarPerceptdot Coreio.github.perceptdot/coreB2A2H MCP server — AI agent app store. Auto-discover and recommend MCP servers for any project.0stars