mcpware/cross-code-organizer

Cross-Code Organizer (formerly Claude Code Organizer): cross-harness config dashboard for Claude Code, Codex CLI, MCP servers, skills, memories, agents, sessions, security scanning, context budget, and backups.

What it solves

Cross-Code Organizer (CCO) addresses the fragmentation and invisibility of configuration files for AI coding tools. Many AI harnesses (like Claude Code, Codex CLI, OpenCode, and DeepSeek Harness) store settings, memories, skills, and MCP servers in hidden directories, making it difficult for users to see what is actually active, identify duplicate configurations that waste context tokens, or detect malicious tool descriptions in MCP servers.

How it works

CCO acts as a local control panel that uses harness-specific adapters to normalize and visualize the configurations of different AI coding tools. It scans local directories (such as ~/.claude/ and ~/.codex/) to provide a unified dashboard. Key features include:

  • Effective Context Mapping: Shows exactly what is loaded per project, accounting for inheritance and precedence rules.
  • Context Budgeting: Calculates token usage for pre-loaded items to help users reclaim context window space.
  • Harness Doctor: Audits the setup for hygiene issues, such as duplicate identities or stale sessions, and provides safe auto-repair.
  • Security Scanner: Analyzes MCP server tool definitions using static patterns and deobfuscation techniques to detect prompt injection or poisoning. It also includes a research preview of an activation-probe scanner that uses a local sensor model to detect hidden activations associated with malicious capabilities.

Who it’s for

Developers using AI coding harnesses like Claude Code, Codex CLI, OpenCode, or DeepSeek Harness who want to manage their AI environment, optimize token usage, and secure their MCP integrations.

Highlights

  • Cross-Harness Support: Unified management for Claude Code, Codex CLI, OpenCode, and DSH.
  • Token Hygiene: Per-item token counts and @import expansion to visualize context window consumption.
  • Security First: Static and activation-probe scanning for MCP tool poisoning with direct navigation to findings.
  • Zero-Install: Can be launched quickly via npx.
  • Config Management: In-dashboard Markdown editing, bulk operations, and reversible moves/deletes.

Related

  • Project
    mcpdotdirect/evm-mcp-server

    An MCP server that enables AI agents to interact with 60+ EVM-compatible blockchains, providing tools for token transfers, smart contract interaction with automatic ABI fetching, and ENS resolution.

  • Project
    langchain4j/langchain4j-spring

    A collection of Spring Boot starters that simplify the integration of LangChain4j into Java applications, reducing boilerplate for AI capabilities.

  • Project
    tournesol-app/tournesol

    Tournesol is a collaborative platform for identifying public-interest videos to build an open database for AI ethics and recommendation system research.

  • Project
    triton-inference-server/backend

    A framework and set of utilities for creating backends that allow Triton Inference Server to execute models from various deep-learning frameworks or custom C/C++ logic.