← 返回 Skills 市场
doanbactam

harness-engineering

作者 doanbactam · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
440
总下载
0
收藏
2
当前安装
1
版本数
在 OpenClaw 中安装
/install harness-engineering
功能描述
Evolve an existing repository toward Harness Engineering by making it more legible to agents, moving critical knowledge into repo-local artifacts, adding pre...
使用说明 (SKILL.md)

Harness Engineering

Evolve an existing repository toward Harness Engineering without pretending it must be rebuilt as a blank template. Prefer repo-local knowledge, predictable structure, and executable checks over aspirational documentation and hidden context.

Workflow

1. Audit the current repo

  • Read AGENTS.md and the top-level docs under docs/.
  • Identify missing source-of-truth artifacts, dead references, placeholder CI, and stack-specific assumptions.
  • Confirm whether the repo is a real app, a governance layer, or a starter template.
  • Preserve existing product and runtime structure where possible. Do not force a greenfield layout onto a mature codebase.

2. Establish the minimum repo-local artifacts

  • Keep AGENTS.md short and map-like.
  • Add only the missing artifacts that materially improve agent legibility, such as:
    • docs/ARCHITECTURE.md
    • docs/core-beliefs.md
    • docs/quality-score.md
    • docs/observability.md
    • docs/worktrees.md
    • docs/skills.md
    • docs/product-specs/index.md
    • docs/product-specs/template.md
    • docs/exec-plans/index.md
    • docs/exec-plans/template.md
    • docs/exec-plans/tech-debt-tracker.md
    • docs/generated/README.md
  • Match the artifact set to the repository's maturity. A small service may need fewer documents than a large multi-domain product.
  • If generated artifacts are referenced but not yet produced, add explicit placeholders rather than leaving silent gaps.

3. Replace policy-only claims with enforcement

  • Use the repo's existing stack where needed, but prefer bun for lightweight governance scripts when that does not conflict with the project.
  • Add executable checks for required files, stale placeholders, broken references, and obvious contract drift.
  • Wire CI to run the checks for every push and pull request.
  • Do not leave comments like "run lint later" as the only quality gate.

4. Preserve generality

  • Preserve real project names, domains, and runtime details when the repository is already in active use.
  • Remove only accidental one-off assumptions, stale tribal knowledge, and undocumented conventions.
  • Write docs so a future agent can continue work from the repository itself.
  • Prefer incremental hardening over broad rewrites.

5. Make the repo legible to agents

  • Keep one clear location for each category of knowledge.
  • Put product intent in docs/product-specs/.
  • Put multi-step delivery work in docs/exec-plans/.
  • Put detailed external library notes in docs/references/.
  • Prefer explicit directory conventions over clever local variation.

6. Verify the template

  • Run bun run template:check or the project equivalent.
  • Fix all hard failures before stopping.
  • Treat warnings as backlog only if they are genuinely optional.
  • If the repository is not a git checkout, note that you cannot report git status.

Editing rules

  • Use imperative language in docs.
  • Prefer concise checklists over essays.
  • Keep examples aligned with bun.
  • Avoid claiming that a metric, score, or automation is "automatic" unless code actually updates it.
  • Add placeholder files when the template references an artifact that new projects will generate later.

Reference

Read references/checklist.md when you need a compact audit list while editing or reviewing a template repo.

Deliverables

  • Updated repo-local docs
  • Missing placeholder artifacts filled in
  • Executable validation or governance checks
  • CI wired to run real checks
  • A short summary of what now supports Harness Engineering and what still depends on hidden context
安全使用建议
This skill is internally consistent and does what it says: audit a repo and add docs/checks. Before using it, ensure your agent runtime has any local tools you expect (e.g., bun) or edit the instructions to use your project's tooling; review all proposed repository edits, CI changes, and governance scripts before merging; and if you prefer to prevent autonomous edits, disable implicit invocation or require manual approval so the agent cannot modify repos/CI without a human in the loop. Note: README contains an example local path (C:\Users\kisde...) — harmless example text but verify no accidental personal secrets are included before publishing.
功能分析
Type: OpenClaw Skill Name: harness-engineering Version: 1.0.0 The 'harness-engineering' skill bundle is designed to guide an AI agent in improving repository structure, documentation, and automated governance. It focuses on creating repo-local artifacts (e.g., AGENTS.md, ARCHITECTURE.md) and implementing executable checks using 'bun' to ensure repository legibility for agents. There is no evidence of malicious intent, data exfiltration, or unauthorized execution; the instructions in SKILL.md and references/checklist.md are consistent with standard software engineering and repository maintenance practices.
能力评估
Purpose & Capability
The name/description match the actual instructions: audit a repo, add repo-local docs, create executable checks, and wire CI. There are no unrelated requested credentials, binaries, or config paths that don't belong to this task.
Instruction Scope
SKILL.md stays within scope: it instructs the agent to read repo docs, add artifacts, create checks, and run a local validation command. Two points to note: (1) it recommends using the bun tool for lightweight governance scripts and suggests running `bun run template:check` — but the skill does not declare bun as required, so runtime may fail if bun is not present; (2) the instructions expect the agent to modify the repository and wire CI pipelines, which is appropriate for the purpose but is an action that should be reviewed by a human before merging or pushing.
Install Mechanism
This is an instruction-only skill with no install spec and no code files — lowest-risk installation footprint. Nothing will be downloaded or executed by the platform as part of an install.
Credentials
The skill requests no environment variables, credentials, or config paths. The guidance to use local tooling and CI is proportional to its goal and does not ask for unrelated secrets.
Persistence & Privilege
The skill is not always-on and declares default model invocation behavior. agents/openai.yaml sets policy.allow_implicit_invocation: true, which permits implicit/autonomous invocation. That is coherent with an agent-facing skill but means the skill can be invoked by the model without an explicit user button press; combine this with the fact the skill edits repos and wires CI — you may want to control invocation or require human review of changes.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install harness-engineering
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /harness-engineering 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release. Introduces the Harness Engineering skill for evolving repositories with agent-friendly structure and executable governance. - Documents a step-by-step workflow for auditing and improving repositories. - Emphasizes repo-local artifacts, incremental add-ons, and clear structure (not greenfield rewrites). - Recommends replacing policy-only docs with executable checks and wiring them into CI. - Provides detailed artifact templates and editing rules for legibility and agent continuity. - Supplies audit and delivery checklists to guide repository improvement.
元数据
Slug harness-engineering
版本 1.0.0
许可证 MIT-0
累计安装 2
当前安装数 2
历史版本数 1
常见问题

harness-engineering 是什么?

Evolve an existing repository toward Harness Engineering by making it more legible to agents, moving critical knowledge into repo-local artifacts, adding pre... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 440 次。

如何安装 harness-engineering?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install harness-engineering」即可一键安装,无需额外配置。

harness-engineering 是免费的吗?

是的,harness-engineering 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

harness-engineering 支持哪些平台?

harness-engineering 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 harness-engineering?

由 doanbactam(@doanbactam)开发并维护,当前版本 v1.0.0。

💬 留言讨论