commits-to-issues
Reads a repository's Git commit history and groups commits into proposed units of work, using path mappings, conventional-commit scopes, or touched folders, with timing gaps and commit metadata influencing the plan.
Creates and closes GitHub issues from generated plans, applies labels derived from commit types, adds comments linking the associated commits, and avoids duplicates by checking existing marked issues.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
commits-to-issues
Generate GitHub issues from your git history. Your commits already record what you shipped; this turns them into closed GitHub issues, grouped into real units of work, each with a checklist of what changed and a comment linking every commit.
A CLI and an MCP server for Claude Code, Codex, and Cursor. You see the whole plan first; nothing is written until you say yes. Run it again and it skips what already exists.
Use it when the issue tracker fell behind the code: a solo project that never had issues, a sprint nobody logged, or a repo you are about to open-source and want a readable history for.
Quick start
Requires Python 3.10+ and a GitHub token (gh auth login is enough).
uvx commits-to-issues --since 2026-09-0142 commits in ~/code/shop → 7 issues for octo/shop
id worked area kind commits title
1 2026-09-01→2026-09-03 orders enhancement 6 Add order export
2 2026-09-04 auth bug 2 Refresh expired sessions
3 2026-09-08→2026-09-10 billing enhancement 9 Support annual plans
...
Nothing was written to GitHub. Re-run with --apply to create and close these issues.Then create them:
uvx commits-to-issues --since 2026-09-01 --applyEach issue gets the from-commits label, plus bug or enhancement from conventional
commit types, and is closed as completed with a comment listing its commits.
Related MCP server: agentic-sdlc-mcp
How commits are grouped
A path map you give it:
--module apps/web/=Web --module api/=APIOtherwise the conventional-commit scope:
fix(auth): ...→authOtherwise the top-level folder the commit touched most
A group splits when work pauses for more than three days (--max-gap-days). The
title comes from the largest commit in the group. Weekend work is dated on the next
business day.
Other options: --until, --author, --github OWNER/REPO (defaults to the origin
remote), --skip 2,5, and --include-uncommitted, which adds one open issue for work in progress.
With an AI agent (MCP)
The same flow is an MCP server for Claude Code, Codex, Cursor, and other clients:
claude mcp add commits-to-issues -- uvx commits-to-issues mcpplan_issues_from_commitsis read-only. It returns the proposed issues and aplan_id.create_issues_from_commitstakes theplan_id, optional newtitlesandskipids, and defaults todry_run=true.
Try this prompt:
Read this repository's git history since September 1 and propose GitHub issues
with plan_issues_from_commits. Show me the plan as a table. Don't create anything until I approve.Safe to retry
Every issue carries a hidden marker derived from its first commit. Before writing,
commits-to-issues lists the repo's from-commits issues and skips any it already made,
so a re-run, or a retry after a network error, never duplicates an issue. Writes are
spaced out and back off on GitHub rate limits.
Limits
GitHub does not let you backdate an issue, so the real dates appear in the issue body ("Worked: 2026-09-08 → 2026-09-10"), not as its created date.
Commit messages are the only evidence. No tests are run.
Also by the author
plane-workflow-mcp does the same for Plane, cloud or self-hosted, plus a planning workflow for agents.
MIT licensed.
This server cannot be deployed
Maintenance
Related MCP Connectors
Assign Linear, Jira or GitHub issues to an AI coding agent team that plans, builds, reviews, tests.
Operate GitHub from your AI assistant: repos, PRs, issues, Actions. Previews and read-only mode.
Plan Salesforce deploys, open pull requests and trigger pipelines from your AI client.
Connect AI assistants to GitHub - manage repos, issues, PRs, and workflows through natural language.
Related MCP Servers
AlicenseNot gradedqualityCmaintenanceEnables code reviews, implementation planning, and pull request generation for AI agents in IDEs like Cursor and Windsurf.1MIT- AlicenseAqualityAmaintenanceEnables AI coding agents to orchestrate the full software development lifecycle on GitHub, including planning, issue creation, code review, security triage, and release readiness checks.1377 npm1MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to automate GitHub repository management, issue tracking, and commits using natural language.6 npmApache 2.0
- AlicenseNot gradedqualityAmaintenanceTurns GitHub Projects v2 into an agentic task substrate where AI coding agents register, atomically claim and complete tasks, exchange heartbeats, review one another's work products, and operate within token budgets. Humans get AI-powered PRD generation, task breakdown, sprint planning, roadmap creation, and issue triage, with all state stored natively in GitHub issues, projects, and comments so no external infrastructure is needed.453 npmMIT