Report Abuse

Basic Information

Patchwork is a self-hosted CLI agent designed to automate common software development chores by orchestrating LLM-assisted workflows called patchflows. It provides a modular system of reusable Steps (atomic actions), Prompt Templates (LLM prompts with placeholders) and Patchflows (compositions of steps and prompts) to perform tasks such as PR review, vulnerability remediation, dependency upgrades, docstring and README generation, and issue resolution. The CLI runs patchflows locally, inside IDEs, or as part of CI/CD. It supports multiple LLM backends including OpenAI-compatible endpoints and local model servers, and offers optional dependency groups for security scanning, retrieval-augmented generation, and notifications. The project also supplies a template repository for default configuration and prompts and invites contributors to add new steps and patchflows.

Links

Categorization

App Details

Features
Modular building blocks: Steps for atomic actions, Prompt Templates for customizable LLM queries, and Patchflows to combine them into reusable automations. Predefined patchflows include GenerateDocstring, AutoFix, PRReview, GenerateREADME, DependencyUpgrade and ResolveIssue. A CLI interface runs patchflows with simple key=value arguments and supports configuration files. Supports OpenAI-compatible endpoints, hosted managed keys, and local model servers via common runtimes. Optional dependency groups enable security scanning and RAG features. Includes a separate template repository containing default configs and prompts. Designed for extensibility with documentation and a chat assistant to help create new steps and patchflows. Can be integrated into CI/CD pipelines and run against repositories in the current directory.
Use Cases
Patchwork reduces repetitive developer work by automating review, bug fixing, vulnerability patching, dependency updates and documentation generation using LLMs. Teams can run safety- and QA-oriented patchflows in their local environment or CI to produce PRs, comments, patches and summaries automatically, freeing engineers to focus on higher-level design and implementation. Its modular steps and prompt templates make behaviors transparent and customizable, enabling secure self-hosting and use of organization-approved models. Optional security and RAG integrations add scanning and repository-aware issue resolution. The CLI and config-driven design make it easy to integrate into existing developer workflows and pipelines, and the template repo plus contribution guides enable teams to extend or create new automated patchflows.

Please fill the required fields*