Top Picks

Open-source orchestrator for running coding agents like Claude Code, Codex, and OpenCode in parallel on a laptop or VPS, with task queues, worktrees, checks, and a web cockpit.

A native Mac app that lets you type @@ in any editable field to talk to your AI agent, with screen, clipboard, and Finder context. It also includes TabTab autocomplete, clipboard history, screenshots, and screen recording.

Open-source structured memory for AI agents. Ingot stores tool results and documents in SQL-friendly form, with optional embeddings, MCP access, and a local quick start for running the stack.

A local sandbox for running LLM-generated code from Python or Node with no cloud, no daemon, and strong isolation controls.

A free AI usage and quota monitor for Codex, Claude Code, and Gemini CLI. It surfaces activity, remaining limits, reset timing, and trend views across desktop and mobile.

Anthropic’s documentation page on how to prompt Claude Opus 5.5, with guidance for effort calibration, agentic runs, progress updates, multi-app workflows, refusals, pasted text, and visual inputs.

Weekly tracking for whether ChatGPT, Perplexity, Google AI Overviews, Gemini, and Claude cite your site, name your brand, or leave you out. The page shows pricing, the report structure, example outputs, and how the system works.

A Windows tray app for monitoring Claude Code usage limits, remaining context, and history, with alerts, charts, and local storage.

DrivingBench is a benchmark that tests frontier LLMs on real-world driving tasks in a Toyota Corolla with a cone course, leaderboard, run traces, and replay videos.

A local proxy that inspects LLM requests and attachments for secrets before they reach the model, with redaction, rejection modes, and local attachment decoding/OCR.

Evidence-based grading for Claude Code skills, agents, slash commands, MCP configs, and plugins, with file-and-line findings, scorecards, badges, and a local/CI workflow.

A curated map of AI safety and frontier AI entry points, covering research programs, fellowships, residencies, policy tracks, courses, directories, and job boards for people outside traditional academic or lab pipelines.

A Claude Code plugin and MCP-based project-state tool that keeps decisions, constraints, rejected paths, and other session context versioned in git so AI sessions can resume consistently.

Errand is an open-source macOS app for working with persistent AI teammates that each have their own cloud computer, conversation history, and files. The page shows the product purpose, how it works, download availability, and a visible FAQ explaining setup and runtime requirements.

A curated marketplace for AI agent skills, where each skill is reviewed by a person and shows before-and-after prompt evidence. It also exposes a JSON API, OpenAPI spec, and MCP server for browsing and using skills programmatically.

aclif is an open-source Agent CLI framework for building command-line tools that AI agents can use across SaaS platforms. It emphasizes one shared grammar, canonical names, introspection, safety metadata, and multiple runtime modes including shell, tool call, and gateway.

A C++ OpenAI-compatible LLM gateway designed for very low latency, with support for translating requests to providers like Anthropic, Gemini, and Cohere.

Open-source execution runtime for production agents. It records tool and LLM calls as durable steps, supports resume after interruption, and lets teams allow, deny, or approve steps with human oversight.

A macOS menu-bar app that supervises MCP servers so multiple editors and clients can share one running server per profile, with credentials stored in Keychain and tool calls recorded.

Vehla is a native macOS command center with on-device AI, clipboard intelligence, window management, dictation, and local-first workflows. It combines launcher-style actions, clipboard recall, shelves, workspace tools, and model connectivity in one app.

Stroq is a local action firewall for AI coding agents. It scans tool output, marks risky sessions, blocks dangerous follow-up actions, and keeps a hash-chained audit log, with support for Claude Code, Cursor, Codex, Copilot CLI, Windsurf, OpenClaw, and MCP clients.

Open-source agentic workspace for teams that combines apps, chat tools, integrations, models, and agents in one local-first environment with shared memory.

A headless Roslyn/MSBuild indexer for C# that gives coding agents compiler-resolved usage, reference, inheritance, and override data across projects.

Bastiontrace is a Python library for analyzing AI agent tool-call traces to locate prompt injections, identify where they landed, and map the resulting blast radius.