harness-execution-layer

Grasp helps you manage multiple projects and parallel coding agents in one visual workflow, so humans can understand faster and ship more.

This is the corresponding code for the O'Reilly book: Harness Engineering

Let your AI apps use your Mac: a local server and menu bar app that lets Claude, ChatGPT, Codex and other MCP apps work in your Mac's apps, files, terminal and browser, with approvals you control.

The agent harness you steer — structured context, automatic staleness gates, and block-by-block review, all bound to your code. Local-first, MCP-native, open source.

The workflow harness for Codex: typed gates, validated evidence, controlled transitions, repair paths, and inspectable logs for any workflow.

A pixel-art office where AI agents actually do the work.

Give any repo a canonical AI control plane. One prompt, any AI — Codex, Claude, Copilot, Cursor.

A compact, schema-first standard for the boundary between cognition and execution.

Harness-kk-god5.0

End-to-end engineering harness for Claude Code — go from planning to building to reviewing in one workflow. 8 agents with multiple modes. In-built GUI to visualise your PRDs and todos. Ship with AI confidently. MIT.

A full fledged kit to yield 80% better results and cost aware vibe code development. Make the best usage of Claude Code AND Codex, inspired by the heavily used antigravity-kit by vudovn, feel free to experiment, raise issues and contribute.

The open source runtime enforcement for AI agents.

kaipi（开辟）是一个把对话组织成树的极简 coding agent：随时开岔探索，弯路整枝扔掉，结论可嫁接回主干，无需重复地为失败探索付费，高效使用每一个 token！

Open Agent Engine for long-running agents. More agents per GPU. No compaction. Every request rebuilt from a typed state frame so prompts stay within a fixed budget no matter how long the task runs. Built and measured on real software engineering work. AGPL-3.0 or commercial license.

A free VRAM calculator for AI models. Not just for LLMs: text generation, embeddings, vision, multimodal, image diffusion, video, audio, and tabular workloads, across inference, LoRA/QLoRA fine-tuning, and full training. Every calculation runs in the browser. Demo app to simultaneously build frontend harness.

Rest Assured skill pack for designing, implementing, documenting, and reporting API tests in Java and CI workflows.

Local end-to-end testing lab for any MCP server: coding agents role-play real users, drive your tools over multiple turns, score runs, and render/interact with MCP Apps ui:// widgets.

Drop in any model. Get a harness. Offline-first CLI + TUI that detects model type, runs the right harness, serves HTTP, and scaffolds plugins — for tabular, embeddings, LLMs, vision, diffusion, computer use, and more.

Orchestrator for Claude Code agents. Builds software guardrailing Spec-driven-development. It loops over a given to-do list until each ticket is implemented, committed, and checked off. Multiple instances can claim tasks in parallel and are synchronized through the Git repository.