Explore

O(1) prompt footprint for long-horizon agents. The SKILL.state paper (arXiv:2608.26263) reimplemented — eq. 8 reproduces exactly, 82.5x at 60 files. 100% coverage.

Node.js·MIT·★ 2·Few stars·No key

Zenith: a continuous-improvement harness for long-running agent tasks. Turns Claude Code, Codex, or Hermes into a multi-agent mission orchestrator via MCP/ACP.

Skill·Apache-2.0·★ 368·No key

Vibe Check is a tool that provides mentor-like feedback to AI Agents, preventing tunnel-vision, over-engineering and reasoning lock-in for complex and long-horizon agent workflows. KISS your over-eage

Node.js·MIT·★ 501·No key

Local-first, eval-first memory for long-horizon AI agents — no LLM at ingest. Python SDK + MCP server with source-traceable recall, belief revision, selective forgetting, and reproducible benchmarks.

Python·MIT·★ 15·No key

An open, thin, model-agnostic harness for long-horizon autonomous AI research. Built on Goose.

Skill·Apache-2.0·★ 13·No key

Industrial benchmark for AI memory – measuring how accurately LLM and agent memory systems recall long-term context, with uniform, reproducible scoring across datasets and memory products.

Skill·Apache-2.0·★ 9·No key

A semantic, transactional, replayable MCP control plane that lets autonomous AI agents operate Dwarf Fortress as a long-lived civilization. Safe-Rust phase-0B scaffold: MVCC world state, witnessed sem

Skill·MIT·★ 23·No key

Provenance-aware memory for AI agents.

Python·MIT·★ 10·Needs a key

Open-source SuiteCRM MCP server. Connect Claude Desktop, Claude Code, or any AI agent to SuiteCRM in under 5 minutes. OAuth2/OIDC, 24 tools, full observability.

Node.js·MIT·★ 1·Few stars·No key