ai-agent-harness

Awesome list for AI agent harness engineering: tools, patterns, evals, memory, MCP, permissions, observability, and orchestration.

Lightweight AI Agent Harness for agentic coding: let strong models explore while humans steer with minimal specs, checkpoints, approval, validation, and reverse sync.

A work-centered runtime for agentic software engineering. Work persists; agents, context, and graphs assemble around it.

Quality-first harness for autonomous AI coding agents — deterministic gates + a default-FAIL evaluator over a Beads task spine. Claude Code plugin; done is earned, not asserted.

Author-once AI agent harness for Claude Code subagents, AGENTS.md, and native cross-platform coding agent outputs.

Open source AI agent harness & agentic AI runtime in TypeScript. 6-phase state machine, 24 tools, human-in-the-loop permissions, MCP server, and composable agent APIs. Zero dependencies.

Governed Agent (AI Agent Harness) Controlplane — run & assure AI digital employees (Hermes Agent) under ISO 27001 / 42001 / AI Verify. Desktop app (Windows / macOS / Linux), OEM of Hermes One & Hermes Agent.

Open admission layer for LLM agent harnesses: forbidden-path quarantine + cross-iteration escalation monitor. v0.1 heuristic, not full IFC.

AI agent governance harness: baton workflow, fleet LLM routing (Ollama/Claude/OpenRouter), and CI gates for Copilot, Claude Code, and Codex.

Harness your AI agent with a systematic workflow to draft accurate, analyst-ready RFI responses (Gartner MQ, Forrester Wave, and similar).

The real framework Actian's PM team uses to harness AI agent to draft hallucination-free, technically accurate, analyst-ready RFI responses.

BenchClaw measurement and evidence layer for AI-agent framework benchmarks

The seven-agent delivery workflow Kromatic uses to ship client software with Claude Code — PM, orchestrator, architect, design, dev, QA, and analytics, with human gates between each.