RLRepoLoot

Open-source coding agents and tools for Claude Code, Cursor and Codex

Coding agents write, refactor and test code from a prompt. The open-source layer around them — skills, harnesses, review hooks, session tooling — is where most of the practical leverage now sits, and it is far less crowded than the assistants themselves.

Most professional developers now run two or three of these at once: a terminal agent for multi-file work, an editor assistant for daily typing, and something running in the background on a branch. That split is what creates the opportunity — the assistants are commercial, but the connective tissue around them is open source and largely unbuilt.

The projects here are the connective tissue: skill libraries, review gates, session recorders, harnesses that turn a chat transcript into a reproducible task. Each one is described by the problem it removes, because that is what decides whether it is worth wiring into a workflow you already have.

Projects in this topic

46 analysed projects match this topic.

Developer Tools
Featured

Developer-first agent tools layer for claude

Skip weeks of custom work by starting from a working agent tools foundation.

agent toolsclaudedeveloper tools
Beginner · AI application
TOOL-76566View details
LLM and Claude Tools
Featured

Practical seo system built around sem

Skip weeks of custom work by starting from a working seo foundation.

seosemgoogle ads
Beginner · Developer library
TOOL-42990View details
Browse the full catalog

Questions people ask

Are these replacements for Claude Code or Cursor?
A few are — there are fully open coding agents in this list. Most are not: they extend an assistant you already pay for, which is usually the cheaper and faster route to a working setup.
What is worth building in this space right now?
The unsolved problems are verification and memory: proving generated code does what was asked, and carrying context between sessions without pasting it back in. Both are thin on the ground and both are things teams will pay for.
Can I use these commercially?
Each project keeps its own licence, and the catalog record names it where the source declares one. Permissive licences dominate the category, but check the file before you build a product on top.

Related collections

Open-Source LLM and Claude Tooling

Open-Source Developer Tools Worth Building On

Open-Source AI Agents to Build On

All collections

Other topics

Agent frameworksMCP toolingRAG and vector searchLocal and self-hosted AILLM inferenceAI code securityBrowser automationTerminal and CLI

Updated: 2026-08-17