Trending

Orbit is a multi-repo Git workspace tool for coding agents. It uses real worktrees and full git history so an agent can read, modify, and commit across several repositories in one unified workspace, with support for Claude Code, Codex, OpenCode, and Qoder.

A two-part self-study course for building a tiny LLM stack from scalar autograd to a working chat assistant, with modules covering transformers, pretraining, instruction tuning, retrieval, tool use, and agent loops.

An AI-powered GitHub Action that reviews pull requests with multi-agent analysis, AST-based line verification, and duplicate suppression.

A multi-tenant MCP server for structured, semantics-aware code refactorings. It supports real refactorings like rename, extract, inline, and organize-imports, with project registration, diff previews, and support for Java, Rust, and TypeScript/JavaScript.

A CLI for searching local agent session history and tracing code changes back to the sessions that produced them.

A CLI and TypeScript helper that lets coding agents turn MCP tool usage into scripts, with typed APIs and a skill for Claude projects.

A macOS feedback layer for coding agents that captures comments on text, screenshots, web elements, and voice, then passes them to agents through MCP with attached context.

Cadre is a local-first Kanban board for orchestrating coding agents, including Claude Code, Codex, Grok, Opencode, and OpenAI-compatible setups. It routes tasks, handles blockers and approvals, and keeps agent runs, diffs, and command control in one workspace.

Open-source self-hosted inference server for agent workloads, with an OpenAI-compatible API, multi-model serving, and production cluster tooling.

TokenSpend tracks AI coding spend and ties it back to shipped work, with GitHub-based PR attribution, team-level reporting, and a live model router with receipts.

A non-destructive compression gateway for AI coding agents that sits between the agent and the LLM API to reduce token usage while keeping original content recoverable on demand.

XSAF is a small, fetch-native TypeScript framework for building secure AI agents. The page shows its core API, agent lifecycle, tool pipeline, MCP support, session handling, and offline testing story.

Open-source CLI that gives AI agents access to web reading, search, and platform-specific retrieval across sites like YouTube, Reddit, GitHub, Twitter/X, Bilibili, Xiaohongshu, and RSS.

ModelPlane is an OpenAI-compatible AI gateway that routes requests across multiple model providers through one endpoint. It includes model groups, fallback and weighted routing, multiple provider support, reasoning output normalization, and usage/billing tracking.

TokenMaxxer is a CLI-based AI usage tracker for developers that reads token counts from tools like Claude Code, Codex, Cursor, and others, then shows spend, model usage, and a public leaderboard. It emphasizes local-only reads and opt-in sharing.

AI forecasting platform with a public track record, API/docs, and integrations for Claude Desktop and Claude Code. The page highlights benchmark results, live market usage, and example forecast types for probability, numeric, date, categorical, and decision questions.

Local-first coding agent that runs on the user’s machine and can work across the browser, desktop, terminal, and editor surfaces. The repo says it verifies fixes by reproducing failures and re-running assertions, with install instructions for Windows, macOS, and Linux.

SKI is a Mac and Windows voice assistant for AI coding agents. It lets developers talk to tools like Claude Code, Cursor, Codex, Gemini CLI, and OpenClaw, and hear replies spoken back locally on-device.

Open-source, local-first usage and cost meter for AI coding agents. It reports usage metadata from tools like Claude Code, Codex, Cursor, Grok Build, Hermes, opencode, Cline, and Copilot CLI, with budgets and warnings and no account or telemetry.

Mirrors is a staging and replay environment for AI agents. It rebuilds the systems an agent calls, replays real sessions against prompts, tools, and model changes, and surfaces regressions before deployment.

NoWreck is a CLI verifier that checks whether AI coding assistant claims match the actual code changes using structural analysis rather than another model judgment.

CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities in code, with repository scanning, change review, finding history, and CI support.

A Python framework for building AI agents as Python objects, with typed methods, stateful fields, and LLM-driven agentic methods. The repo includes installation steps, a quick start, examples, and notes on supported models and safety boundaries.

Hubo is a coding-agent workflow that uses two persistent roles — a work agent and a review agent — to iterate on code changes until review findings are resolved or a human decision is needed. The repo includes installation steps and usage examples for Claude Code, Codex, GitHub Copilot CLI, OpenClaw, and skills.sh.