Browse
44 tools
- Claude Code Analysis · other
82 docs and 15 diagrams mapping every major subsystem of Claude Code's accidentally exposed 512K-line TypeScript source - YOLO classifier, 93% context compaction, prompt-cache layout, 88+ feature flags, the custom React-Fiber terminal renderer.
- Claudraband · typescript
Wraps the real Claude Code TUI with a session lifecycle layer. Resumable non-interactive workflows, HTTP daemon for remote/headless control, ACP server for editor integrations (Zed, Toad). Drives your existing Claude Code install rather than reimplementing it - keeps skills, hooks, MCPs, and approvals intact.
- Garden Skills · javascript
Three carefully-scoped skills: web-design-engineer (with an anti-cliche blocklist that breaks the generic-AI-landing-page loop), gpt-image-2 (80+ templates, three runtime modes including advisor-only fallback), and kb-retriever (layered data_structure.md navigation for bounded local-KB retrieval). Tested across Claude Code, Claude.ai, Cursor, Codex, Gemini, OpenCode.
- mcptube · python
MCP server that turns YouTube videos into a persistent, merging wiki rather than ephemeral vector chunks. Scene-change frame extraction + vision analysis captures slides, code, and diagrams that transcripts miss. 25+ MCP tools, FTS5+LLM hybrid retrieval, version history with source attribution per claim.
- PostTrainBench · python
Benchmark measuring whether Claude Code, Codex CLI, Gemini CLI, and OpenCode can autonomously improve 4 small base models (Qwen3-1.7B/4B, SmolLM3-3B, Gemma-3-4B) on 7 evals (AIME, BFCL, GPQA, GSM8K, HealthBench, HumanEval, Arena Hard) within a single H100 GPU and 10 hours. Includes agent-as-judge anti-reward-hacking and baseline-replacement penalties for tampering.
- Rapid-MLX · python
MLX-native inference engine with OpenAI-compatible API. The novel piece: DeltaNet state snapshots bring prompt caching to non-trimmable architectures (Qwen3.5 hybrids), restoring RNN state in ~0.1ms. 2-5x faster TTFT, native Metal kernels, continuous batching.
- trace-mcp · typescript
MCP server with 138 tools and cross-language framework awareness (58 integrations across 81 languages). Indexes Laravel/Inertia/Vue, Rails/Hotwire, Django/HTMX edges so agents skip re-deriving call graphs. Decision memory links architectural choices to the code they're about. Local-first ONNX embeddings, optional LSP enrichment.
- wanman · typescript
Multi-agent runtime that spawns each Claude Code or Codex agent in its own git worktree and home directory. JSON-RPC subprocess control, task pooling, artifact storage. Solves the share-a-directory failure mode that breaks most multi-agent harnesses.
- agent-of-empires · rust
TUI and web UI for managing parallel Claude Code, OpenCode, Codex, Gemini, Copilot, and Factory Droid sessions. Backed by tmux and git worktrees so each agent gets its own branch.
- AgentBox
One SDK to run Claude Code, Codex, or OpenCode inside Docker, E2B, Modal, Daytona, or Vercel sandboxes - boots each agent's native server (JSON-RPC, HTTP/SSE) instead of using non-interactive --print mode.
- caveman · python
Claude Code skill that rewrites prompts in caveman speech to cut 65% of tokens with no measurable quality loss. Joke premise, real savings.
- claude-memory-compiler · python
Hooks capture Claude Code sessions, the Agent SDK extracts decisions and lessons, and an LLM compiler organizes them into cross-referenced knowledge articles. Memory that grows with the repo.
- cli-to-js · typescript
Wraps any command-line tool as a typed JavaScript API agents can call directly. Saves writing a custom MCP for every CLI you want to expose.
- CubeSandbox
Open-source sandbox runtime for LLM-generated code built on RustVMM and KVM. Targets sub-60ms cold starts with full kernel isolation, designed as a self-hostable replacement for closed E2B-style services.
- dario · typescript
Local proxy that exposes your Claude Max or Pro subscription as an Anthropic-compatible API. No API key, no per-token billing - just point your tools at localhost.
- DeepZero · python
Vulnerability research framework that parses, decompiles, and analyzes Windows kernel drivers for exploitable IOCTLs using AI agents. Sleep through fuzzing campaigns.
- google-docs-mcp · typescript
MCP server for Google Docs that uses pattern-matching search-and-replace (like file editors) instead of character offsets, which LLMs are notoriously bad at. Fixes the broken existing options.
- helixent · typescript
Small TypeScript library for ReAct-style agent loops on the Bun stack. Tools, skills, and a coding-focused harness in a minimal package.
- Keeper · go
Embeddable Go secret store using Argon2id and XChaCha20-Poly1305 by default, with four security levels, audit chains, and crash-safe rotation. Vault when Vault is overkill.
- kmcp · go
Go CLI plus Kubernetes operator that builds, tests, and deploys MCP servers as native k8s workloads. Wraps server lifecycle in standard kubectl-style flows.
- litmus · typescript
TypeScript CLI that runs unit tests against prompts: compare models, check outputs, track cost. Treats prompts as code that needs CI.
- llm-openai-via-codex · python
Simon Willison's plugin for the `llm` CLI that routes calls through an existing OpenAI Codex subscription. Lets you use Codex-tier models from any `llm`-aware tool.
- lunar · typescript
MCP gateway focused on governance and security: policy enforcement, request inspection, and rate-limiting between agents and MCP servers. Sits between the model and the tool surface.
- mcpm.sh · python
Cross-platform CLI to search, install, and configure MCP servers across Claude Desktop, Cursor, Cline, Continue, Goose, and more. Includes router and profile features for routing requests across servers.
- Orca
Stably's next-gen IDE that's built around running coding agents in parallel. First-class support for Claude Code, Codex, Cursor agent, OpenCode, Ghostty, and worktree-based orchestration.
- pentest-ai-agents
Specialized Claude Code subagents that turn the CLI into a pentest assistant: plan engagements, analyze recon, research exploits, build detections, audit STIGs, and write reports.
- prism · python
Python TUI that finds where extra tokens are burned in Claude Code sessions, why sessions fail, and what to fix. Built on Textual, focused on debugging your own usage.
- rqlite · go
Distributed, fault-tolerant SQL database built around SQLite using Raft consensus. v10 adds a built-in web console and CRC32 integrity checks throughout.
- stacklit · go
Go CLI that generates a compact codebase summary (~250 tokens) and auto-configures Claude Code, Cursor, and Aider with it. Replaces the agent's exploration phase with a single primer.
- talk-normal · shell
System prompt that forces any LLM to drop the corporate-overlord cadence and write like a normal person. Strips em-dashes, hedging, and 'in summary' filler.
- ThinkWatch · rust
Rust gateway in front of OpenAI, Anthropic, Gemini, and self-hosted LLMs (plus MCP servers) with RBAC, audit logs, rate limits, and cost tracking. The boring layer enterprises actually need.
- tui-use · typescript
Lets agents interact with programs that expect a human at the keyboard - REPLs, debuggers, TUI apps - things bash pipes cannot reach. Fills the gap between shell and full computer-use.
- vulnhawk · python
Static analysis scanner that finds auth bypass, IDOR, and business logic bugs that Semgrep and CodeQL miss. Ships as a free GitHub Action covering Python, JS/TS, Go, PHP, and Ruby.
- weave · rust
Rust merge driver that resolves false conflicts when independent agents edit the same file, claiming roughly 95% reduction versus line-based merge by working at entity granularity.
- abtop · rust
Rust TUI that monitors Claude Code and Codex sessions: token spend, context window, rate limits, and ports in real time. Like btop but for agent runtimes.
- agents-md
Curated AGENTS.md preset that kills sycophancy, blocks drive-by refactors, and forces verification loops. Synthesizes Karpathy's principles with Cherny's Claude Code workflow.
- awesome-harness-engineering
Curated awesome list for AI agent harness engineering: tools, patterns, evals, memory, MCP, permissions, observability, and orchestration.
- cc-telegram-bridge
Multi-bot, multi-engine Telegram bridge with per-bot personality, budget caps, streaming, session resume, and an Agent Bus for parallel pipelines.
- Codeburn
Interactive terminal dashboard that breaks down where your AI coding tokens actually go. Surfaces the chat-vs-tool-use split most users get wrong.
- design-extract
npx CLI plus Claude Code plugin that extracts colors, typography, spacing, and shadows from a live site into a structured design-token report.
- ESP-Claw
Espressif's chat-coding agent framework for ESP32 devices. Brings tool-calling LLM agents to embedded targets with C-level memory budgets.
- M-flow
Memory layer for agents that surfaces relevant context instead of just similar embeddings. Episodic plus long-term memory backed by a knowledge graph.
- Obscura · rust
Open-source Rust headless browser built for AI agents and scraping. Lower memory and faster cold starts than Chromium-based stacks like Puppeteer and Playwright.
- Qartez MCP · rust
Rust MCP server that pre-indexes a codebase with PageRank, blast-radius, git co-change, and complexity signals. Agents query the graph instead of grepping files.