Individual skill
agentic-harness-engineering
Agent Harness Evolution — 基于 HARNESS.md v1.0 规范的评估、审计与改进工作流。加载后自动检查当前 workspace 结构合规性，引导 evidence-driven 的 Harness 改进。当用户要求做 Harness 审计、初始化 Agent workspace、生成或验证 Change Manifest、对 harness 组件做 evidence-driven 改进，或反复出现同一类失败模式时触发。
skills/agentic-harness-engineering/SKILL.md
Source commit 182ea06e578df7b1d503eca537503c1e4c5a8a58
What was checked
- Matches the observed repository HEAD
- Yes
- Scanner rules are current
- Yes
- Repository scan complete
- No
- Skill scan complete
- Yes
- Observed behavior
- Local scripts
- Last checked
- Nov 10, 2026
These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.
Install this skill
Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.
An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.
Repository context
- Repository stars (not skill ratings)
- 969
- Repository last updated
- Oct 30, 2026
- License
- MIT
Observed agent integration: codex
- engineering Derived category
Files and excerpt
File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.
Browse files
SKILL.mdcore · 7,286 bytesprofiles/codex.yamlasset · 869 bytesprofiles/hermes.yamlasset · 885 bytesprofiles/openclaw.yamlasset · 918 bytesreferences/HARNESS.mddoc · 14,299 bytesreferences/PROJECT_README.mddoc · 7,127 bytesreferences/analysis/README.mddoc · 702 bytesreferences/analysis/detail-template.mddoc · 414 bytesreferences/analysis/overview-template.mddoc · 843 bytesreferences/change-manifest-schema.jsonasset · 5,335 bytesreferences/directory-template.mddoc · 2,015 bytesreferences/docs/results-format.mddoc · 2,403 bytesreferences/docs/rollback-strategy.mddoc · 4,556 bytesreferences/examples/minimal-workspace/AGENTS.mddoc · 161 bytesreferences/examples/minimal-workspace/MEMORY.mddoc · 60 bytesreferences/examples/minimal-workspace/SOUL.mddoc · 66 bytesreferences/examples/minimal-workspace/manifests/example_manifest.jsonasset · 755 bytesreferences/examples/minimal-workspace/tool_descriptions/example.tool.yamlasset · 230 bytesreferences/examples/minimal-workspace/tools/example_tool.pyruns · 69 bytesreferences/examples/trace-sample/analysis-overview.mddoc · 1,477 bytesreferences/examples/trace-sample/task-t042-search-truncation.mddoc · 2,060 bytesscripts/ci.shruns · 4,342 bytesscripts/generate_manifest.pyruns · 7,790 bytesscripts/init_harness.pyruns · 4,731 bytesscripts/validate_harness.pyruns · 12,383 bytesscripts/verify_manifest.pyruns · 5,988 bytes
Read excerpt
# AHE Skill — Agent Harness Evolution
> 将 [AHE（Agentic Harness Engineering）](https://arxiv.org/abs/2604.25850) 论文的核心模式实现为可执行的 Skill。
> 遵循 [HARNESS.md v1.0](references/HARNESS.md) 规范。
---
## 什么时候加载这个 Skill
- **做 Harness Audit** — 想知道当前 workspace 的结构是否符合规范 → 加载 skill，运行 audit 流程
- **改了一个核心组件** — 修改了 system prompt、工具、middleware 等 → 加载 skill，生成 Change Manifest
- **想验证上次修改** — 之前生成了 Change Manifest 但还没验证 → 加载 skill，运行 verify 流程
- **Harness 反复出同一个问题** — 同一个失败模式出现 2+ 次 → 加载 skill，做 evidence-driven 改进
- **搭建新 Agent** — 初始化一个新的 Agent workspace → 加载 skill，用 HARNESS.md 模板初始化