Skip to main content
Glama
Sign In
RepoControl MCP

RepoControl MCP

by johannschiffer
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

RepoControl MCP (Lite Edition)

License: MIT Protocol: MCP Python: >=3.10

RepoControl MCP is a secure, least-privilege Model Context Protocol server that enables AI agents (Claude Desktop, Cursor, LangGraph, CrewAI) to manage GitHub issues, inspect discussions, create pull requests, and audit CI check-runs with strict safety guards.

🎯 The Problem

Developers are increasingly delegating code refactoring and bug-fixing to autonomous agents:

  1. Destructive Force Pushes: Giving an LLM a blanket repo token risks accidental force pushes to main, corrupting commit histories or deleting production branches.

  2. Abandoned Failing PRs: AI agents often push commits, open a PR, and immediately conclude the prompt—leaving failing GitHub Actions builds and broken test suites unaddressed.

  3. Missing Audit Trails: No cryptographic evidence ledger tracking which commits and diffs were proposed by which agent model.

RepoControl acts as an intelligent control plane between your autonomous agents and the GitHub API, enforcing safety guardrails and verifying builds before work is marked complete.

Related MCP server: ReviewGuard MCP

✨ Features (Lite Edition)

  • 🛡️ Branch Safety Guards: Built-in protection preventing destructive operations against main, master, prod, release.

  • 📝 Issue & PR Lifecycle: Safely create issues, fetch discussions, and open pull requests.

  • 🚦 CI Status Auditing: Inspect GitHub Actions check-runs and overall build conclusions directly via MCP tools.

  • ⚡ Zero-Config Protocol: Standard JSON-RPC 2.0 stdio server compatible with Claude Desktop and Cursor.

🚀 Quickstart

1. Installation

git clone https://github.com/johannschiffer/repocontrol-mcp.git
cd repocontrol-mcp
pip install -e .

2. Configure in Claude Desktop or Cursor

Add to your claude_desktop_config.json or Cursor MCP settings:

{
  "mcpServers": {
    "repocontrol": {
      "command": "repocontrol",
      "args": ["run"],
      "env": {
        "GITHUB_TOKEN": "ghp_your_github_token_here"
      }
    }
  }
}

3. CLI Inspection

# Check CI status of any commit or branch
repocontrol ci-check owner/repo feature-branch

🛠️ MCP Tools Reference

Tool

Parameters

Description

create_issue

repo, title, body, labels

Opens an issue in the target repository

get_issue_comments

repo, issue_number, limit

Retrieves discussion comments on an issue/PR

create_pull_request

repo, title, head, base, body

Opens a PR with branch collision guards

list_pull_requests

repo, state, limit

Lists open or closed pull requests

check_ci_status

repo, ref

Audits GitHub Actions check-runs and conclusion

is_branch_protected

branch

Validates if a branch is covered by safety guards

🚀 Upgrade to RepoControl Pro

Need enterprise-grade autonomous CI recovery and security governance?

The RepoControl Pro Edition adds:

  • 🔍 CI/CD Log Analyzer & Auto-Fix: Automatically extracts compiler errors, test failures, and stack traces from failing GitHub Actions runs so the LLM can self-correct without human triage.

  • 🔒 Least-Privilege Token Scoper & Command Filter: Restricts agent capabilities strictly to authorized operations and blocks destructive API endpoints.

  • 📜 Cryptographic Evidence Audit Ledger: SQLite WAL ledger recording SHA-256 hashes of every commit, diff, and test report, with automated Markdown attestation for PRs.

  • 🐳 Production Docker & Compose: Containerized sidecar setup for cloud agent runners.

  • 🧪 100% Passing Pytest Suite ready for enterprise CI/CD.

👉 Get RepoControl Pro on Gumroad ($14.99 USD)

📄 License

MIT © 2026 Johann Schiffer

Available Tools

6 tools
check_ci_statusB

Inspect GitHub Actions check-runs and overall CI status for a branch or commit SHA.

Inputs: ref (required), repo (required).

create_issueC

Create a new issue in a GitHub repository.

Inputs: body (required), repo (required), title (required), labels.

create_pull_requestB

Open a Pull Request with built-in branch safety validation.

Inputs: head (required), repo (required), title (required), base, +1 more.

get_issue_commentsC

Retrieve recent discussion comments on an issue or pull request.

Inputs: repo (required), issue_number (required), limit.

is_branch_protectedB

Check if a branch name is considered protected by safety rules.

Inputs: branch (required).

list_pull_requestsC

List pull requests in repository filtered by state.

Inputs: repo (required), limit, state.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    Claude Code, inside the Godot 4 editorFlockbay puts an AI agent in the real Godot 4 editor: it writes GDScript, runs your game and fixes what breaks. Free, or bring your own Claude Code or Codex.flockbay.comAd

    TDQS

    B3.4/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool has a clearly distinct target action: creating issues, reading comments, creating/listing pull requests, checking CI, and checking branch protection. The only slight overlap is get_issue_comments covering issue or PR comments, but the intent remains clear.

    Naming Consistency5/5

    All tool names use consistent snake_case with a predictable verb_noun or is_property pattern. The naming is readable and uniform across the set.

    Tool Count5/5

    Six tools is well-scoped for a focused repository-control server. Each tool appears to earn its place without obvious redundancy.

    Completeness3/5

    The set covers issue creation/comments, PR creation/listing, CI status, and branch protection checks, but lacks obvious lifecycle operations such as updating/closing issues, merging PRs, or reviewing PRs. These gaps may force agents to work around missing operations.

    Resources

    Related MCP Connectors

    Related MCP Servers