Maintained officially by GitHub, the GitHub MCP server exposes issues, pull requests, code search, repository content, Actions runs, and more to LLM agents. Authentication is via GitHub OAuth or PAT, scoped to whatever your token can do. Use cases: PR-triage agents, code-review preprocessors, issue-summary digests, "find the function that does X" search-and-fetch.
Pricing
Free · Official (MIT) · GitHub plan applies
Developer Consensus: Pros
Official GitHub maintenance — auth, rate limiting, and versioning are first-class52× mentioned
Code search returns symbol-level context, not just file paths41× mentioned
PR-summary mode reads the diff and writes the description33× mentioned
Actions integration lets agents trigger workflows and read logs24× mentioned
Common Friction Points
Rate limits on personal tokens get hit fast on big monorepos18× mentioned
No native handling for sub-issues / nested workflows11× mentioned
Setup with GitHub Enterprise Server needs an extra config flag7× mentioned
Code search across private orgs requires fine-grained PAT scope review6× mentioned
Verified Peer Reviews
P
@pr_agent
Engineering Lead · TypeScript · Mid
Verified
Auto-PR-summaries that are actually good.
GitHub MCP plus Claude reads the diff, scans the changed files for related context, and writes the PR description. Reviewer time drops 30% because the description tells them what the PR is for instead of just what changed.
I
@issue_triage
Open Source Maintainer · Python · Solo
Verified
Issue triage that does not require me opening every issue.
My agent reads incoming issues, classifies into bug/feature/question, tags appropriately, and writes a first-pass reply for me to approve. 200-issue backlog became a 15-minute review per day.
R
@rate_pain
Platform Engineer · Go · Enterprise
Verified
Move to GitHub App tokens, not PATs.
PAT rate limits hurt on a 5,000-repo monorepo org. Switching to a GitHub App with finer-grained installation tokens fixed it.
Every review on this page is verified through GitHub OAuth and weighted by reviewer credibility, use-case match, and conflict-of-interest disclosure. Aggregate scores combine with recency decay so rankings reflect current reality.
Read full methodology →