{"id":51234717,"url":"https://github.com/catesandrew/mcp-agent-bridge","last_synced_at":"2026-06-28T19:03:28.574Z","repository":{"id":358317808,"uuid":"1231006481","full_name":"catesandrew/mcp-agent-bridge","owner":"catesandrew","description":"Bridge AI coding agents (Claude Code, Codex, Copilot) as MCP servers for cross-agent code review and collaboration","archived":false,"fork":false,"pushed_at":"2026-05-16T18:27:54.000Z","size":319,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-16T20:36:16.658Z","etag":null,"topics":["ai-agents","claude","code-review","codex","copilot","mcp","model-context-protocol","typescript"],"latest_commit_sha":null,"homepage":"https://catesandrew.github.io/mcp-agent-bridge/","language":"TypeScript","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/catesandrew.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-06T14:31:37.000Z","updated_at":"2026-05-16T18:27:58.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/catesandrew/mcp-agent-bridge","commit_stats":null,"previous_names":["catesandrew/mcp-agent-bridge"],"tags_count":3,"template":false,"template_full_name":null,"purl":"pkg:github/catesandrew/mcp-agent-bridge","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/catesandrew%2Fmcp-agent-bridge","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/catesandrew%2Fmcp-agent-bridge/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/catesandrew%2Fmcp-agent-bridge/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/catesandrew%2Fmcp-agent-bridge/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/catesandrew","download_url":"https://codeload.github.com/catesandrew/mcp-agent-bridge/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/catesandrew%2Fmcp-agent-bridge/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34900378,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-28T02:00:05.809Z","response_time":54,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-agents","claude","code-review","codex","copilot","mcp","model-context-protocol","typescript"],"created_at":"2026-06-28T19:03:28.037Z","updated_at":"2026-06-28T19:03:28.569Z","avatar_url":"https://github.com/catesandrew.png","language":"TypeScript","funding_links":[],"categories":[],"sub_categories":[],"readme":"# MCP Agent Bridge\n\n[![CI](https://github.com/catesandrew/mcp-agent-bridge/actions/workflows/ci.yml/badge.svg)](https://github.com/catesandrew/mcp-agent-bridge/actions/workflows/ci.yml)\n[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)\n[![Node](https://img.shields.io/badge/node-%3E%3D20-brightgreen)](https://nodejs.org)\n\nBridge AI coding agents (Claude Code, OpenAI Codex, GitHub Copilot) as [MCP](https://modelcontextprotocol.io) servers so they can call each other for code reviews, analysis, and cross-validation.\n\n**[Documentation](https://catesandrew.github.io/mcp-agent-bridge/)**\n\n## Why?\n\nAI coding agents are powerful individually, but they each have blind spots. MCP Agent Bridge lets you:\n\n- **Get a second opinion** -- Send your diff to a different AI model for review\n- **Cross-validate** -- Compare responses from Claude, Codex, and Copilot on the same code\n- **Structured reviews** -- Get typed JSON output with verdict, issues by severity, and suggestions\n- **Share instances** -- Run each agent once as a background service, shared by all clients\n\n## Quick Start\n\n```bash\n# Clone and build\ngit clone https://github.com/catesandrew/mcp-agent-bridge.git\ncd mcp-agent-bridge\nbun install \u0026\u0026 bun run build:exe\n\n# Start the Claude bridge on port 8940\nnpx mcp-proxy --port 8940 -- ./exe/claude-mcp-server\n```\n\nConfigure your MCP client (`.mcp.json`):\n\n```json\n{\n  \"mcpServers\": {\n    \"claude_reviewer\": {\n      \"type\": \"streamable-http\",\n      \"url\": \"http://localhost:8940/mcp\"\n    }\n  }\n}\n```\n\nThen ask your agent: *\"Use the claude_reviewer MCP to review this diff.\"*\n\n## Architecture\n\nEach bridge runs as a singleton background service behind an HTTP proxy. Multiple clients share one instance.\n\n```\n┌─────────────┐     HTTP/SSE      ┌───────────┐     stdio     ┌──────────────────┐\n│  MCP Client │ ───────────────► │ mcp-proxy │ ────────────► │  claude -p       │\n│  (any agent)│    :8940/mcp     │           │               │  codex exec      │\n└─────────────┘                  └───────────┘               │  copilot -p      │\n                                                             └──────────────────┘\n```\n\n| Server | Port | Core Tools | Resume Tools |\n|--------|------|------------|--------------|\n| Claude | 8940 | `review`, `ask`, `code_review` | 16 career tools |\n| Codex | 8941 | `codex`, `code_review`, `codex_reply` | 16 career tools |\n| Copilot | 8945 | `ask`, `code_review` | 16 career tools |\n\n## Resume \u0026 Career Tools\n\nAll three servers include 16 resume and career tools powered by embedded skill methodologies. No network fetch at runtime — the expert logic is compiled in.\n\n| Tool | Purpose |\n|------|---------|\n| `career_fact_extractor` | Extract a fact-ID-tagged career database from any source material |\n| `cover_letter_generator` | Personalized cover letters with match analysis and talking points |\n| `creative_portfolio_resume` | ATS + designed resume for creative professionals |\n| `executive_resume_writer` | C-suite, VP, and Director-level transformation-story resumes |\n| `interview_prep_generator` | STAR-method prep with predicted questions ranked by probability |\n| `job_description_analyzer` | Match scoring, keyword extraction, and red flag detection |\n| `linkedin_profile_optimizer` | Algorithm-aware LinkedIn profile rewrite |\n| `portfolio_case_study_writer` | Six-section case studies for design, engineering, PM, and marketing |\n| `reference_list_builder` | Reference strategy, formatting, and briefing email templates |\n| `resume_ats_optimizer` | Keyword matching and ATS compliance fixes |\n| `resume_bullet_writer` | Transform duty bullets into achievement bullets using X-Y-Z formula |\n| `resume_formatter` | ATS-safe formatting audit and repair |\n| `resume_quantifier` | Add metrics to every bullet using estimation methodology |\n| `resume_section_builder` | Build targeted sections by career stage |\n| `resume_tailor` | Tailor a master resume to a specific job posting |\n| `recruiter_first_screen_simulation` | Simulate a skeptical hiring manager's 45-second resume screen |\n| `resume_version_manager` | Organize and track multiple resume versions with naming conventions |\n| `tech_resume_optimizer` | Optimize for software engineering, data, DevOps, and technical PM roles |\n\nSee the [Resume Tools Guide](https://catesandrew.github.io/mcp-agent-bridge/docs/guides/resume-tools) and [SKILLS.md](SKILLS.md) for full input schemas and usage examples.\n\n## Structured Review Output\n\nThe `review` and `code_review` tools return predictable JSON:\n\n```json\n{\n  \"verdict\": \"NEEDS_REVISION\",\n  \"issues\": [\n    {\n      \"severity\": \"critical\",\n      \"description\": \"SQL injection via string interpolation\",\n      \"recommendation\": \"Use parameterized queries\"\n    }\n  ],\n  \"suggestions\": [\n    \"Add rate limiting to the login endpoint\"\n  ]\n}\n```\n\n## Installation\n\n### Pre-built Binaries\n\nDownload from [GitHub Releases](https://github.com/catesandrew/mcp-agent-bridge/releases) for macOS (ARM64/x64), Linux (x64/ARM64), and Windows (x64).\n\n### From Source\n\n```bash\nbun install\nbun run build:exe    # Standalone binaries in exe/\n./install.sh         # Copy to ~/.local/bin\n```\n\n### Background Services\n\n```bash\n# macOS -- LaunchAgents\n./install.sh --launchd --load\n\n# Windows -- NSSM services\n.\\examples\\windows\\install.ps1 -HttpService\n\n# Linux -- systemd (see docs)\n```\n\nSee the [Deployment Guide](https://catesandrew.github.io/mcp-agent-bridge/docs/deployment/macos) for details.\n\n## Configuration\n\nAll servers are configured via environment variables:\n\n| Variable | Default | Description |\n|----------|---------|-------------|\n| `CLAUDE_REVIEW_MODEL` | `opus` | Model for Claude reviews |\n| `CLAUDE_MCP_HTTP_PORT` | `8940` | Claude HTTP port |\n| `CODEX_MCP_HTTP_PORT` | `8941` | Codex HTTP port |\n| `COPILOT_MCP_HTTP_PORT` | `8945` | Copilot HTTP port |\n\nSee [all environment variables](https://catesandrew.github.io/mcp-agent-bridge/docs/configuration/environment-variables) in the docs.\n\n## Testing\n\n```bash\nbun run test        # 36 tests across 5 suites\nbun run lint        # TypeScript type checking\n```\n\n## Security\n\n- All proxies bind to `127.0.0.1` (localhost only)\n- Read-only tool allowlist by default (`Read`, `Grep`, `Glob`, `LS`)\n- Working directory validation against allowed roots\n- Input size limits (500KB) and process timeouts (5 min)\n\n## Prerequisites\n\n- [Bun](https://bun.sh/) \u003e= 1.0 (build)\n- Node.js \u003e= 20 (runtime)\n- At least one agent CLI installed and authenticated:\n  [Claude](https://docs.anthropic.com/en/docs/claude-code) |\n  [Codex](https://github.com/openai/codex) |\n  [Copilot](https://githubnext.com/projects/copilot-cli)\n- [`mcp-proxy`](https://github.com/nicholasgasior/mcp-proxy) (for HTTP mode)\n\n## Contributing\n\nSee [CONTRIBUTING.md](CONTRIBUTING.md) for development setup and guidelines.\n\n## License\n\nLicensed under the Apache License, Version 2.0. See [LICENSE](LICENSE) for details.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcatesandrew%2Fmcp-agent-bridge","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fcatesandrew%2Fmcp-agent-bridge","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fcatesandrew%2Fmcp-agent-bridge/lists"}