github Not archived

china-qijizhifeng/agentic-harness-engineering

Official AHE code — Agentic Harness Engineering: observability-driven automatic evolution of coding-agent harnesses (concurrent w/ meta-harness). NexAU-AHE reaches 84.7% ± 2.1 pass@1 on Terminal-Bench 2 (GPT-5.5). Lifts GPT-5.4 69.7→77.0% over 10 iters, beats Codex/ACE/Training-Free GRPO; frozen harness transfers to SWE-bench-Verified.

Quick read

Stars
893
Forks
98
Open issues
9
Commits
47

Activity and growth

Latest capture 2026-09-18 03:06

Stars · last 7 days
No history
Commits · last 7 days
No history
Stars since tracking
+316
Stored snapshots
5

Classification

Technology stack

Metadata

Language
Python
License
MIT
Default branch
main
Created
2026-04-23
First commit
2026-04-26
Last pushed
2026-08-03
GitHub updated
2026-09-17
Last synced
2026-09-18 03:06
Stack scanned
2026-09-18 03:06
Archived
No

Archive status reflects the last catalog sync, not a maintenance assessment. Check recent releases, issue responses, and fit for your project with the repository shortlisting checklist.

AI development signals

2 paths

Agent instructions and tool configuration found in this repository.

Growth history

Tracked growth

5 observed captures since 2026-06-16. Observed captures are shown by default.

Stars from first capture +316

Chart data

Observed captures only

Time horizon

All tracked data

Custom date range

Stars history

Observed snapshots

Commits history

Observed snapshots

Similar repositories

Nearest indexed repositories by embedding similarity.

raphaelchristi/harness-evolver

Automated harness evolution for AI agents. A Claude Code plugin that iteratively optimizes system prompts, routing, retrieval, and orchestration code using full-trace counterfactual diagnosis. Based on Meta-Harness (Lee et al., 2026).

56 stars
Python 1 awesome list

EvolvingAgentsLabs/evolving-agents

A director agent that learns how your team works: it plans by walking your CLAUDE.md and memory as linked statements, delegates to Claude Code or local models, and keeps each lesson as a gated, versioned patch. Measured: 0.929 vs 0.589.

454 stars
TypeScript 0 awesome lists

future-agi/future-agi

Open-source, end-to-end platform for evaluating, observing, and improving LLM and AI agent applications. Tracing · Evals · Simulations · Datasets · Gateway · Guardrails. Self-hostable. Apache 2.0.

1,943 stars
Python 5 awesome lists

epicsagas/epic-harness

Multi-tool AI agent harness — 22 skills, self-evolving engine, unified memory, autonomous spec-to-PR pipeline. Works with Claude Code, Codex, Cursor, OpenCode, and Cline.

19 stars
Rust 1 awesome list

revfactory/harness

A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.

8,926 stars
HTML 3 awesome lists