Node-RED nodes for Cloud File Management

AI agent framework for plan-first development workflows with approval-based execution. Multi-language support for TypeScript, Python, Go, Rust and more.

The repository-installable agent skills live in [`skills/`](./skills/): `assurancespec-authoring` creates and refines deterministic proposals, while `assurancespec-work` executes reviewed obligations without acquiring admission, verification, completion,

OpenCode-style AI SDK Core adapter for Khala Code.

Sails P2P Trading SDK — the single typed client a wallet/fintech imports to reach every Sails Protocol module (OpenIdentity, OpenLiquidity, OpenP2P, OpenSettlement, OpenReputation) over HTTP/WebSocket. See docs/SDK_GUIDE.md.

Effect-native Recursive Language Model engine: symbolic environment, declarative programs, ModelMap/RlmMap, Tier D + Tier S.

OSS, local-first, runtime-agnostic autonomous-QA runner. Drive any web target with a bring-your-own OpenAI-compatible model (no login), record a playable video + Playwright trace + screenshots, and distill the session into a committed e2e test. Khala / Op

The Effect-native **agent harness contract**. One versioned adapter contract drives a third-party coding-agent runtime (Codex, Claude Code, an ACP peer, a managed sandbox) behind a uniform surface with durable, cursor-exact turn suspend and continue.

Coffrify MCP server — gives Claude/Cursor/Windsurf agents direct access to Coffrify : 207+ tools spanning transfers, webhooks, audit, API keys, analytics, branding, domains, folders, collections, members, notifications, GDPR, status, changelog, alerts, te

Vendor-neutral identity verifier for AI agents

Vet a bounty before you point an AI agent at it. Detects prompt-injection attacks aimed at agents and repos that farm free work.

Khala terminal client with interactive and headless CLI modes

Polis CLI — bring your own agent to the AXL-native intelligence network. File sourced signals, archive to 0G, and earn USDC when your work clears review.

Official TypeScript SDK for SBO3L — the cryptographically verifiable trust layer for autonomous AI agents.

Effect-native typed model programs, immutable compilation artifacts, bounded optimization, and receipted prediction.

Canonical in-repo ATIF (Agent Trajectory Interchange Format, ATIF-v1.7) schema, validators, and public-safety tripwire shared by trace producers (qa-runner) and consumers (the trace ingest API).

Zero-config decentralized agent-to-agent communication. Agents npm install, connect, and talk — no servers, no accounts, no surveillance. Built on libp2p + IPFS with encrypted P2P mesh networking, viral content pinning, and automatic peer discovery.

Deploy vertical UniswapX solvers in npm install

Authenticated MCP access to the OpenAgents World AI worker marketplace