📅 Book a 30-min Demo📞 Call/text (571) 293-0242
Connector · Source code hosting and collaboration

GitHub MCP Server & Agent Skill

Last updated

ibl.ai publishes a free GitHub agent skill: a SKILL.md that teaches an AI agent to read and write GitHub data and expose it as an MCP tool. Download it, connect your own GitHub credentials, and run the agent on ibl.ai Agentic OS, where you own all the code and the data.

GitHub holds an organization's Git repositories, issues, pull requests and reviews, Actions workflows and runs, code security alerts, discussions, projects and organization membership.

What is the GitHub API, and what can an agent read from it?

REST endpoints by resource (repos, issues, pulls, actions, orgs) versioned with the X-GitHub-Api-Version header, and a GraphQL API at https://api.github.com/graphql.

  • API: GitHub REST API and GraphQL API (Mixed)
  • Access: Public, self-serve
  • Base URL: https://api.github.com/
  • Authentication: GitHub App (installation access token); Fine-grained personal access token; OAuth app (authorization code); Personal access token (classic)

Any account can create a personal access token or register a GitHub App or OAuth app. Organization owners choose which repositories a GitHub App can access and can require approval for fine-grained tokens.

GitHub Apps use fine-grained permissions and short-lived tokens, acting as the app or for a user. Fine-grained PATs are limited to one owner and chosen repositories; classic PATs use broad scopes.

What can an agent read from GitHub?

  • Issues — GET /repos/{owner}/{repo}/issues.
  • Pull requests — GET /repos/{owner}/{repo}/pulls, /pulls/{pull_number} and /pulls/{pull_number}/files (up to 3,000 files).
  • Repository contents — Read repositories, branches and file contents via the REST or GraphQL API.
  • Actions and security — Workflow runs, code scanning and Dependabot alerts (toolsets on the GitHub MCP Server).

What can an agent write to GitHub?

  • Create issue — POST /repos/{owner}/{repo}/issues; fine-grained permission Issues write.
  • Update issue — PATCH /repos/{owner}/{repo}/issues/{issue_number} to edit, label, assign or close.
  • Create pull request — Open a pull request from a head branch to a base branch; drafts supported.
  • Merge pull request — PUT /repos/{owner}/{repo}/pulls/{pull_number}/merge with merge, squash or rebase. Confirm with the user first.

Limits: 5,000 requests/hour authenticated (15,000 for Enterprise Cloud apps), 60/hour unauthenticated. Secondary limits: 100 concurrent requests, 900 points/min per REST endpoint, 80 content-creating requests/min. 403 or 429 when exceeded.

Repositories hold source code, which is trade-secret intellectual property, and can contain committed secrets or personal data in issues and commits.

How do I connect GitHub to an AI agent over MCP?

GitHub publishes an official MCP server: GitHub MCP Server.

Maintained by GitHub. Remote at https://api.githubcopilot.com/mcp/ or local Docker image; OAuth or PAT; --toolsets and --read-only flags limit what an agent can do.

You can register that server on ibl.ai as it is, or use the SKILL.md below when you want the agent to call the GitHub REST API and GraphQL API directly.

On ibl.ai an MCP connection is three records, and a working integration always needs all three:

  1. Register the server — POST …/mcp-servers/ with its URL, a transport (sse, websocket or streamable_http) and an auth_type of none, token or oauth2.
  2. Create a connection that holds the credential, scoped to the whole organization (platform), to a single agent, or to each user (user).
  3. Enable it on the agent — mcp-tool in tool_slugs and the server id in mcp_servers. Both lists are replaced, not merged, so read the settings first.

GitHub supports per-user OAuth, so the server can use auth_type=oauth2 with auth_scope=user: each person connects their own GitHub account in the chat the first time the agent needs it.

The field reference is in MCP server connections, and the /iblai-api-agent-mcp skill runs all three steps from a coding agent.

What does the GitHub agent skill do?

The GitHub skill is one SKILL.md file named iblai-connect-github. It gives an agent the auth conventions, what to read, which writes need a person's confirmation, the ibl.ai MCP wiring, example prompts and guardrails.

Download the GitHub SKILL.md

Save it as .claude/skills/iblai-connect-github/SKILL.md, or in your agent's skills folder. The ibl.ai platform skills it calls come from one install:

$ npx skills add iblai/vibe --all

This connector skill is a starting point maintained on ibl.ai, not part of the iblai/vibe repository.

It was checked against GitHub's documentation on October 20, 2026. Test it against your own GitHub setup before you rely on it.

What can I ask a GitHub agent to do?

  • “List open pull requests in acme/api that are waiting on my review”
  • “Create an issue in acme/web for the broken checkout button and label it bug”
  • “Summarize the changes in pull request 482 and flag risky files”
  • “Which Dependabot alerts are open across our critical repositories?”

How do I run a GitHub agent on ibl.ai?

Agentic OS is where the agent runs on ibl.ai: it holds the GitHub MCP connection, the model the agent uses, its instructions and its conversation history.

On ibl.ai you own all the code and the data, run it model-agnostic across any LLM, and pay with no per-seat pricing — so you can deploy anywhere, from your own cloud to a fully air-gapped network.

Which ibl.ai agents work with GitHub data?

Which ibl.ai platform skills does a GitHub agent use?

What else should I know before connecting GitHub to an AI agent?

Is there an official GitHub MCP server?

Yes. GitHub publishes GitHub MCP Server.

You can register it on ibl.ai as an MCP server, or use the SKILL.md when you want the agent to call the GitHub REST API and GraphQL API directly.

Who can get GitHub API credentials?

Any account can create a personal access token or register a GitHub App or OAuth app. Organization owners choose which repositories a GitHub App can access and can require approval for fine-grained tokens.

Does GitHub data leave my infrastructure when an ibl.ai agent uses it?

On ibl.ai, you own all the code and the data. Self-hosted in your cloud or data center, the agent, its MCP connections and its conversation logs stay inside your perimeter.

Because the platform is model-agnostic, you can point the agent at a model you host, so GitHub records need not reach a third-party model provider.

Is the GitHub agent skill free to use?

Yes. The SKILL.md is a free download with no sign-up. It assumes you already have GitHub credentials, and your agreement with GitHub still governs how you use its API.

Which other data sources can an AI agent connect to alongside GitHub?

Which other Enterprise systems have a connector?

Every connector, grouped by sector, is on the connectors hub.