Explore
MCP servers and skills: search, filter by where they run, and compare.
conPACTst0nebridge
Local
MCP servers and skills: search, filter by where they run, and compare.
CONtext comPACT coordinates agent-initiated context compaction at verified scope closure—after the final answer and a durable checkpoint. It compacts the live context without clearing the visible conv
PythonMIT★ 3Few starsNo key
Universal long-term project memory for AI coding agents — Claude Code, Cursor, Windsurf, Cline + 4 more. TDD rules, code graph, spec-driven dev, cross-session recall.
SkillMIT★ 31No key
An experimental framework for building collaborative coding agents that design and implement software through structured debate and consensus.
PythonApache-2.0★ 4Few starsNo key
Substrate-based context compaction for Claude Code. Inspectable signal weights, local-only, MCP-ready. Alternative to /compact + hook-based capture (claude-mem, etc.).
PythonMIT★ 5No key
Context health watchdog and verified session handoffs for Claude Code and Codex
SkillMIT★ 1Few starsNo key
Multi-agent context compaction and secure tool-output vault | 多 Agent 上下文压缩与省 Token 工具集
PythonMITFew starsNo key
MCP file cache for AI coding agents. Re-reads cost ~40 tokens instead of the whole file; changed files come back as diffs. Works with Claude Code, Cursor, Codex and VS Code.
PythonMIT★ 3Few starsNo key
AI-driven CAD workbench and coding agent for Windows, evolved from illusion-agent: shared canvas for design discussion, headless 3D preview, SolidWorks-connected real-machine modeling with live visual
PythonMIT★ 2Few starsNo key
Loop-aware MCP server for your own ComfyUI: build → run → LOOK at the pixels → critique → fix. The ratchet (best-so-far + ledger) is durable server-side state, so it survives context compaction. 32 to
PythonMIT★ 3Few starsNo key
AI copilot for TouchDesigner — DeepSeek v4. Standalone .tox (95 tools) + Claude Code MCP (110 tools, incl. td_ingest_url web ingestion). Activity log, OCR sidecar, tool approval gates, trace viewer, v
PythonMIT★ 6No key
A local-first skill to save tokens and keep sessions clutter-free.
SkillMIT★ 7No key
Terminal AI coding agent built from scratch — ReAct loop, multi-LLM providers, 5-layer permissions, MCP, context compaction, subagents & teams. 716 tests.
PythonMIT★ 1Few starsNo key
A personal, self-developing AI agent you run yourself: its own app window on macOS, Linux and Windows, a web app on your phone, Telegram if you want it. Real tools (shell, browser, git, MCP), projects
PythonMIT★ 2Few starsNo key
One project record for rules, decisions, evidence, and action boundaries — for every supported AI coding tool; actions stay constrained, verifiable, auditable.
SkillMIT★ 1Few starsNo key
Specialized-agent harness that lets a lightweight local reasoning model (ornith:9b via Ollama) carry out long, autonomous tasks. Inspired by Claude Code managed subagents.
SkillMIT★ 2Few starsNo key
MCP for Claude Code self-management (hot restart, compaction)
PythonMIT★ 2Few starsNo key
轻量级终端coding agent KKCode — Terminal AI coding agent with two-layer context compaction, MCP lazy-loading, multi-agent team & OS sandbox
PythonMIT★ 1Few starsNo key
MCP server that gives Claude hands, eyes, and a diary for Pokémon — Game Boy, GBA, and a real Switch 🦭
PythonMITFew starsNo key
Every Claude Code session writes down what it did. Prism is the part that reads it back
SkillMITFew starsNo key
Who stands behind it
