← 返回 Skills 市场
159
总下载
0
收藏
2
当前安装
1
版本数
在 OpenClaw 中安装
/install using-harness
功能描述
Force Harness mode for task decomposition, routing, package-first execution, and verifier-first completion evidence.
安全使用建议
This skill is a coherent, low-technical-risk set of workflow rules (no code, no creds). Before enabling it for agents, confirm what your agent will do to satisfy vague items like TASK_ROUTING_MATRIX and Verifier evidence (will it call external project systems, create accounts, or request sensitive files?). Test in a sandbox, confirm verifier roles/processes exist in your org, and decide whether to permit autonomous runs for this policy or require manual approval when the agent attempts package creation or dispatch.
功能分析
Type: OpenClaw Skill
Name: using-harness
Version: 1.0.0
The skill bundle defines a structured workflow ('Harness mode') for task decomposition, routing, and verification. The instructions in SKILL.md are purely procedural, focusing on governance and process management without any executable code, network requests, or attempts to access sensitive system resources.
能力评估
Purpose & Capability
The name/description (enforce a harness-style workflow) matches the SKILL.md content: procedural rules about locking task objects, routing, packaging, and verifier evidence. The skill declares no binaries, env vars, or installs — which is proportionate for a policy/instruction-only skill.
Instruction Scope
The SKILL.md is entirely procedural and does not instruct the agent to read files, access environment variables, or call external endpoints. However, the instructions are high-level and refer to external artifacts (TASK_ROUTING_MATRIX, Task/Verifier Package) without defining how to obtain them; that vagueness gives the agent operational discretion (it may query external systems or ask for context) — something to be aware of but not inherently malicious.
Install Mechanism
No install spec and no code files are present. This instruction-only format minimizes disk footprint and supply-chain risk.
Credentials
The skill requests no environment variables, credentials, or config paths. There is no apparent need for elevated access given the skill's stated purpose.
Persistence & Privilege
always is false and the skill does not request persistent presence or modify other skills/system settings. Autonomous invocation is allowed by platform default but is not elevated by this skill.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install using-harness - 安装完成后，直接呼叫该 Skill 的名称或使用
/using-harness触发 - 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 publish
元数据
常见问题
using-harness 是什么？
Force Harness mode for task decomposition, routing, package-first execution, and verifier-first completion evidence. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 159 次。
如何安装 using-harness？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install using-harness」即可一键安装，无需额外配置。
using-harness 是免费的吗？
是的，using-harness 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
using-harness 支持哪些平台？
using-harness 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 using-harness？
由 tasteed（@villiamsl）开发并维护，当前版本 v1.0.0。
推荐 Skills