mcpField
C
codehost-remote

codehost-remote

Local
github.com/CloudCorpRecordsVersion Control0Updated 2w ago

About codehost-remote

codehost-remote is a CLI and MCP server for forwarding a pinned set of GitHub repository, issue, and pull request tools to GitHub’s hosted MCP endpoint. It is for clients that need these GitHub operations without a local REST client.

Key features

  1. 01Repository and branch operations
  2. 02File content and update operations
  3. 03Repository and code search
  4. 04Issue and pull request operations
  5. 05Forwards JSON-RPC calls and SSE responses

Use cases

  1. 01Search GitHub repositories or code
  2. 02Create and update repository files
  3. 03Review and manage issues
  4. 04Create pull requests

MCP support

Exposes 14 GitHub tools through an MCP server and forwards calls to GitHub’s hosted remote MCP endpoint. Calls use a GitHub personal access token.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"cloudcorprecords-codehost-remote": {
"command": "npx",
"args": [
"-y",
"@reneturcios/codehost-remote"
]
}
}
}
Run commandnpx -y @reneturcios/codehost-remote

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@reneturcios/codehost-remote0.1.0
ocighcr.io/cloudcorprecords/codehost-remote:0.1.0

Tools

14

Saved tool list · updated 1w ago

ParameterTypeRequiredDescription
namestringYesName for the new repository.
privatebooleanNoWhether the new repository is private (true) or public (false).
autoInitbooleanNoWhether to initialize the new repository with a README file.
descriptionstringNoA short description of the new repository.

Details

From the project README

codehost-remote

codehost-remote gives you the familiar GitHub tool surface — repositories, branches, files, search, issues and pull requests — without a local REST client. Every call is forwarded as a JSON-RPC tools/call to GitHub's hosted remote MCP endpoint at https://api.githubcopilot.com/mcp/, and the hosted answer, whether plain JSON or a streamed SSE response, is what you get back.

That makes it the lightweight option: no GitHub API versioning to track locally, and the hosted service does the work. The trade-off is that you depend on the hosted endpoint's availability and behaviour.

Only a core subset is exposed here: 14 tools. The hosted endpoint actually serves a much larger, dynamic set of toolsets (40+ tools); this package deliberately pins the repos / issues / pull-requests core and never claims the full hosted surface.

Set GITHUB_PERSONAL_ACCESS_TOKEN; it is sent as a Bearer token. GITHUB_REMOTE_MCP_URL overrides the endpoint (useful for testing).

Install

npx @reneturcios/codehost-remote get_issue '{"owner":"octocat","repo":"hello-world","issue_number":1}'
docker run -i -e GITHUB_PERSONAL_ACCESS_TOKEN ghcr.io/cloudcorprecords/codehost-remote

CLI

Every tool is a subcommand taking one JSON argument object:

codehost-remote search_repositories '{"query":"mcp language:typescript"}'
codehost-remote create_issue '{"owner":"me","repo":"proj","title":"Bug: startup crash"}'
codehost-remote --version

Exit codes: 0 success, 1 error, 2 usage error.

MCP server

npx -y @reneturcios/codehost-remote mcp
{
  "mcpServers": {
    "codehost-remote": {
      "command": "npx",
      "args": ["-y", "@reneturcios/codehost-remote", "mcp"],
      "env": { "GITHUB_PERSONAL_ACCESS_TOKEN": "your-token" }
    }
  }
}
Tools

create_repository, fork_repository, create_branch, get_file_contents, create_or_update_file, push_files, search_repositories, search_code, list_issues, get_issue, create_issue, add_issue_comment, list_pull_requests, create_pull_request.

A missing token is reported as a tool error naming GITHUB_PERSONAL_ACCESS_TOKEN.

Development

npm install
npm run build
npm test

Tests run entirely on fake-fetch fixtures, including plain-JSON and SSE upstream bodies; no network is used.

License

MIT — see LICENSE.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.