Skip to main content
Glama
itsocialist

Claude Code Connector MCP

by itsocialist

Claude Code Connector MCP

An MCP server that bridges Claude Desktop, Claude Code CLI, and Claude Code for VS Code, enabling seamless developer workflows across all three interfaces.

Overview

This MCP server allows developers to:

  • Plan and research in Claude Desktop

  • Push specs and documents directly to project directories

  • Invoke Claude Code CLI from Desktop conversations

  • Share state and context across all Claude interfaces

Project Status

Current Phase: Planning
Target Developer: antigravity
Priority: P0

Documentation

  • PRODUCT_SPEC.md - Complete product specification with all features, tools, resources, prompts, and slash commands

  • ROADMAP.md - Implementation phases and timeline

  • ARCHITECTURE.md - Technical architecture (to be created)

Quick Start (Post-Implementation)

Installation

npm install -g @ciq/claude-code-connector-mcp

Configuration

Add to your Claude Desktop config:

{
  "mcpServers": {
    "claude-code": {
      "command": "claude-code-connector",
      "args": ["--projects-config", "~/.claude/projects.json"]
    }
  }
}

Usage

/projects                          # List registered projects
/project myapp                     # Set active project
/save docs/spec.md                 # Save to project
/implement create auth middleware  # Invoke Claude Code
/checkpoint completed auth         # Create checkpoint

Development Phases

Phase 1: MVP (Core File Operations)

  • Project registration and management

  • File read/write to project directories

  • Basic slash commands: /projects, /project, /save

  • Estimated: 1 week

Phase 2: Claude Code Integration

  • CLI invocation with task delegation

  • Streaming status updates

  • /implement command

  • Estimated: 1 week

Phase 3: State Management

  • Checkpoint system

  • Session state persistence

  • Cross-interface context sharing

  • Estimated: 1 week

Phase 4: Advanced Features

  • Git integration

  • MCP resources and prompts

  • Performance optimization

  • Estimated: 1 week

Key Features

MCP Tools (7 total)

  1. register_project - Register codebase directory

  2. list_projects - Show all registered projects

  3. write_to_project - Write file to project

  4. read_from_project - Read file from project

  5. invoke_claude_code - Execute Claude Code CLI task

  6. create_checkpoint - Save conversation state

  7. get_project_status - Get project overview

MCP Resources (3 total)

  • claude-code://{projectId}/files - Project file tree

  • claude-code://{projectId}/session - Session state

  • claude-code://{projectId}/docs - Project documentation

MCP Prompts (4 total)

  • plan-and-implement - Full workflow guide

  • save-to-project - Quick save artifacts

  • continue-in-cli - Handoff to CLI

  • project-status - Project overview

Slash Commands (6 total)

  • /projects - List projects

  • /project {id} - Set active project

  • /save {path} - Quick save

  • /implement {task} - Invoke Claude Code

  • /checkpoint {summary} - Create checkpoint

  • /status - Project status

Architecture

Components

  • MCP Server: Node.js/TypeScript server implementing MCP protocol

  • Project Manager: Handles project registration and file operations

  • CLI Invoker: Spawns and manages Claude Code processes

  • State Manager: Handles session state and checkpoints

  • Resource Provider: Exposes project data as MCP resources

Data Storage

  • ~/.claude/projects.json - Project registry

  • {project}/.claude/session_state.json - Per-project session state

Success Metrics

MVP Success

  • Developer can register project from Desktop

  • Developer can write specs to project directory

  • Developer can read project files

  • Basic commands work reliably

Full Feature Success

  • Developer can invoke Claude Code from Desktop

  • State persists across interfaces

  • Streaming updates work

  • <500ms response time for typical operations

Contributing

This is a CIQ internal project. Development by antigravity.

Development Setup

git clone [repository]
cd claude-code-connector-mcp
npm install
npm run build
npm link  # For local testing

Testing

npm test              # Run unit tests
npm run test:integration  # Run integration tests
npm run test:e2e     # Run end-to-end tests

License

Proprietary - CIQ Inc.

Support

For issues or questions, contact Brian (Product Manager) or antigravity (Developer).

Roadmap

See ROADMAP.md for detailed implementation plan.

Open Questions

  1. Does Claude Code CLI support --task and --context flags?

  2. What's the best streaming protocol (SSE, websockets, polling)?

  3. Should MCP handle git operations directly?

  4. How should VS Code extension consume state files?

Status: Awaiting technical discovery from antigravity.

Available Tools

4 tools
list_projectsC

List all registered projects with status

ParametersJSON Schema
NameRequiredDescriptionDefault
includeInactiveNoInclude inactive projects

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It states it lists projects 'with status', hinting at behavioral output, but doesn't disclose key traits like pagination, sorting, default behavior (e.g., active-only unless specified), authentication needs, or rate limits. For a list tool with zero annotation coverage, this is inadequate.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, efficient sentence with zero waste. It's front-loaded with the core action and resource, making it easy to parse quickly. Every word contributes directly to the purpose without redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given no annotations and no output schema, the description is incomplete for a list tool. It lacks details on return format (e.g., structure of projects, what 'status' entails), behavioral constraints, or error handling. With low complexity but missing output info, it should provide more context for effective use.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, with one parameter ('includeInactive') fully documented in the schema. The description adds no parameter-specific information beyond implying status inclusion, which the schema already covers. Baseline 3 is appropriate as the schema handles parameter semantics effectively.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the verb ('List') and resource ('all registered projects'), specifying the scope with 'with status'. It distinguishes from siblings like 'register_project' (create) and 'write_to_project' (modify), though not explicitly from 'read_from_project' (which might imply reading content vs. listing metadata).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance is provided on when to use this tool versus alternatives. It doesn't mention if it's for overviews, filtering, or how it differs from 'read_from_project' (e.g., listing metadata vs. reading project data). The description implies a general listing function but lacks context on prerequisites or exclusions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

read_from_projectC

Read file content from registered project

ParametersJSON Schema
NameRequiredDescriptionDefault
projectIdYesID of registered project
filePathYesRelative path within project
startLineNoStart reading from line N (1-indexed)
endLineNoStop reading at line N (1-indexed)

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries full burden but lacks behavioral details. It doesn't disclose permissions required, rate limits, error handling, or output format (e.g., text string, binary data). 'Read' implies a safe operation, but without annotations, more context is needed for safe invocation.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, clear sentence with zero waste. It's front-loaded with the core purpose and appropriately sized for the tool's complexity, making it easy to parse quickly.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given no annotations and no output schema, the description is incomplete. It doesn't explain what is returned (e.g., file content as text, error if file not found), behavioral traits, or usage context. For a read operation with 4 parameters, more information is needed for reliable agent use.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so parameters are well-documented in the schema. The description adds no additional meaning beyond implying file content reading, which aligns with the schema. Baseline 3 is appropriate as the schema handles parameter semantics adequately.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description 'Read file content from registered project' clearly states the action (read) and target (file content from registered project), distinguishing it from siblings like list_projects, register_project, and write_to_project. However, it doesn't specify what type of content is read (e.g., text, binary) or how it's returned, keeping it from a perfect score.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus alternatives. It doesn't mention prerequisites (e.g., project must be registered), exclusions, or comparisons to siblings like write_to_project for modifications. Usage is implied by the name but not explicitly stated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

register_projectC

Register a project directory for Claude Code Connector access

ParametersJSON Schema
NameRequiredDescriptionDefault
nameYesHuman-readable project name
rootPathYesAbsolute path to project root
idNoOptional unique ID (auto-generated if not provided)
specPathsNoRelative paths for spec/doc storage

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It states the tool 'registers' a project directory, implying a write/mutation operation, but doesn't disclose critical traits like whether this requires specific permissions, if it's idempotent, what happens on duplicate registration, or any rate limits. The mention of 'access' hints at enabling future operations but lacks details on effects or constraints.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, efficient sentence that front-loads the core purpose ('register a project directory') without unnecessary words. Every part earns its place by specifying the resource and context ('for Claude Code Connector access'), making it appropriately sized for the tool's complexity.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool involves mutation (registration) with no annotations and no output schema, the description is incomplete. It doesn't explain what the registration enables, potential side effects, error conditions, or return values. For a 4-parameter tool that likely modifies system state, more context on behavior and outcomes is needed to guide an agent effectively.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already documents all 4 parameters thoroughly (e.g., 'name' as human-readable, 'rootPath' as absolute path). The description adds no additional meaning beyond what the schema provides, such as explaining how 'specPaths' relate to 'Claude Code Connector' functionality. Baseline 3 is appropriate when schema does the heavy lifting.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the action ('register') and resource ('project directory'), specifying it's for 'Claude Code Connector access'. It distinguishes from siblings like 'list_projects' (listing vs registering) and 'read_from_project/write_to_project' (accessing vs registering). However, it doesn't explicitly differentiate from potential overlapping tools beyond the given siblings.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus alternatives like 'list_projects' (which might show existing registrations) or prerequisites for registration. It mentions 'Claude Code Connector access' but doesn't clarify if this is for initial setup, re-registration, or specific contexts.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

write_to_projectC

Write content to a file in registered project directory

ParametersJSON Schema
NameRequiredDescriptionDefault
projectIdYesID of registered project
filePathYesRelative path within project
contentYesFile content to write
createDirsNoCreate parent directories if they don't exist
overwriteNoOverwrite if file exists

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries full burden for behavioral disclosure. It states it's a write operation but doesn't mention permission requirements, error conditions (e.g., invalid paths), whether it's idempotent, or what happens on success/failure. The description is minimal and misses important behavioral context for a file write tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, clear sentence with zero wasted words. It's front-loaded with the core action and target, making it immediately understandable. Every word earns its place in this concise formulation.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a write operation tool with 5 parameters, no annotations, and no output schema, the description is incomplete. It doesn't explain what 'registered project' means (context from sibling 'register_project'), doesn't mention the tool's effect on existing files, and provides no information about return values or error handling. The description leaves significant gaps for proper tool usage.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so all parameters are documented in the schema. The description adds no additional parameter semantics beyond what's in the schema (like explaining what 'registered project' means or format expectations). Baseline 3 is appropriate since the schema does the heavy lifting.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the action ('write content') and target ('to a file in registered project directory'), which is specific and actionable. It distinguishes from siblings like 'read_from_project' by specifying a write operation, but doesn't explicitly differentiate from other potential write operations beyond the project context.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus alternatives. It doesn't mention prerequisites (like needing a registered project), when not to use it, or how it relates to sibling tools like 'register_project' (which might be required first). Usage context is implied but not explicitly stated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

TDQS

A3.5/5.0
Disambiguation5/5

Each tool has a clearly distinct purpose with no overlap: list_projects enumerates projects, register_project adds new ones, read_from_project retrieves file content, and write_to_project modifies files. The actions (list, register, read, write) and targets (projects vs. files) are well-differentiated, leaving no ambiguity for agent selection.

Naming Consistency5/5

All tools follow a consistent verb_noun pattern with underscores: list_projects, register_project, read_from_project, write_to_project. The naming is uniform and predictable, making it easy for agents to infer functionality from the names alone without deviation or mixed conventions.

Tool Count5/5

With 4 tools, the set is well-scoped for the server's purpose of managing projects and files in the Claude Code Connector. Each tool earns its place by covering essential operations (listing, registering, reading, writing), avoiding bloat or thinness, which is ideal for a focused utility server.

Completeness5/5

The tool surface provides complete CRUD/lifecycle coverage for the domain: register_project (create), list_projects (read/list), read_from_project (read content), and write_to_project (update). There are no obvious gaps for managing projects and files, ensuring agents can handle core workflows without dead ends.

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/itsocialist/claude-code-connector-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server