OpenClaw Automates GitHub Maintenance With Conservative Issue to PR Loops

May 4, 2026

OpenClaw, an open-source self-hosted AI assistant that runs across messaging platforms, released ClawSweeper 0.2.0. This update introduces an autonomous maintenance loop that manages a GitHub issue from start to finish: building code, submitting a guarded pull request (a proposed code change), and performing iterative repairs before automerging.

This release mirrors the industry shift toward Anthropic's autonomous repair loops where background workers handle tedious software maintenance. By splitting operations into four distinct lanes—Review, Apply, Repair, and Commit Review—the bot provides a structured, conservative alternative to the high-volume "bot spam" often seen in automated triage tools.

You can deploy the bot to manage backlogs using a "Read, Write, Propose" flow that ensures no action is taken on raw model output without a markdown report. The tool is open-source, allowing you to run maintenance tasks locally or via GitHub Actions for low-latency repository dispatch.

View the full update on clawsweeper.bot

OpenClaw🦞

@openclawMay 3

ClawSweeper 0.2.0 🦞 The OpenClaw maintenance bot now handles the loop: issue → @clawsweeper fix/build → guarded PR → review → repair → re-review → automerge Still conservative. Much less manual. https://t.co/Z6MW1ftU2m

951.2k

View on X

Still wondering? A few quick answers below.

ClawSweeper is a specialized maintenance bot for OpenClaw repositories that automates GitHub issue triage and pull request management. It acts as a conservative assistant that reviews issues, commits, and pull requests, only taking action when there is strong evidence. It aims to reduce manual maintenance work while keeping repository backlogs clean and actionable for human developers.

The bot follows a multi-step autonomous cycle starting with an issue identification. It then builds a fix, submits a guarded pull request, and enters a review phase. If errors are found, it performs repairs and re-reviews the code before finally automerging the changes. This loop allows the bot to handle the entire lifecycle of a bug fix autonomously.

ClawSweeper organizes its tasks into four distinct lanes: Review, Apply, Repair, and Commit Review. The Review lane handles scheduled issue and pull request triage, while the Apply lane manages guarded mutations like closing items. The Repair lane executes the fix-and-merge loop for opted-in requests, and the Commit Review lane scans code changes on the main branch for findings.

ClawSweeper is an open-source tool that users can self-host on their own hardware. It is designed to work with GitHub repositories using repository dispatch for low-latency reviews or through scheduled GitHub Actions. Because it is part of the OpenClaw ecosystem, it follows a self-hosted model where the user maintains full control over the bot and its rules.

ClawSweeper is considered conservative because it never acts directly on raw model output. Every decision is first recorded as a markdown report in a separate repository for maintainer review. It only proposes closing an issue when specific criteria are met, such as the item being unreproducible or stale, and it never auto-closes items authored by human maintainers.