Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add cilidinezy-commits/doc-harness --skill skill-zhgit clone --depth 1 https://github.com/cilidinezy-commits/doc-harnessWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/cilidinezy-commits/doc-harness/skill-zh)<a href="https://agentmods.dev/skills/cilidinezy-commits/doc-harness/skill-zh"><img src="https://agentmods.dev/badge/skills/cilidinezy-commits/doc-harness/skill-zh/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/cilidinezy-commits/doc-harness/skill-zh"><img src="https://agentmods.dev/badge/skills/cilidinezy-commits/doc-harness/skill-zh.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00191 | $0.02059 |
| Opus 5 | $0.00096 | $0.01030 |
| Sonnet 5 | $0.00038 | $0.00412 |
| Haiku 4.5 | $0.00019 | $0.00206 |
Grade A, and why
doc-harness scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Doc Harness — 基于文档的项目控制系统
Doc Harness 是一套文档管理系统，使任何 AI Agent 或人类协作者能够仅通过阅读文件来理解和恢复项目工作状态——无需任何外部记忆系统。
每个项目创建并维护五个文档：
- CLAUDE.md — 项目入口（概述、恢复链、铁律、操作规则）
- CURRENT_STATUS.md — 活跃状态（车辙/车身/车灯/驾驶手册）
- FILE_INDEX.md — 按类别组织的文件目录
- WORKLOG.md — 永久工作历史（只增不改）
- DOC_HARNESS_SPEC.md — 完整规范（参考文档）
另有两个可选文档，当项目真正积累了相应内容时才创建（详见 spec.md 第13章）：
- PARKING_LOT.md — 暂缓事项，含复活前置条件
- PHILOSOPHY.md — 本项目实践催生的原则
当项目需要与其他项目协调时，可启用一套可选机制（详见 spec.md 第14章）：
- 跨项目 inbox/outbox — 自含式文件消息协议。
inbox/和outbox/目录；带 YAML 前置信息的 Markdown 消息；生命周期unread → read → actioned。完整规范内嵌于 doc-harness 本身，无需外部文档。
命令
/doc-harness init [项目名] [描述]
为新项目初始化 Doc Harness。在当前目录创建全部5个文件。
→ 详见 init.md
/doc-harness check
审计当前项目的文档健康状况并反思工作原则。
→ 详见 check.md
/doc-harness sync [--auto]
将状态文档与现实同步。修复漂移、刷新陈旧字段、注册缺失文件，并可选择触发阶段切换或 WORKLOG 归档。
interactive（默认）：在阶段切换、归档或创建新原则文档之前询问用户。auto：执行修复，无需询问。
→ 详见 sync.md
/doc-harness flush [--auto]
上下文压缩前的紧急保存。包含 sync 的一切功能，加上强制性地将上下文中的重要信息提取到文档中。
flush 分为五个阶段：(A) 同步、(B) 上下文盘点、(C) 写入并注册、(D) 验证、(E) 冲刷标记。阶段 B 和阶段 C 是区分特征——它们扫描智能体当前上下文中仅存于内存的信息，对其分类并写入文件。没有阶段 B 和阶段 C，flush 与 sync 毫无区别，即冲刷失败。
interactive（默认）：在每次重要提取之前询问用户。auto：使用启发式规则对上下文信息进行分类和保存，无需询问。
→ 详见 flush.md。含空扫描报告要求（当无可提取条目时必须产出）。
/doc-harness resume [--auto]
结构化状态恢复：当上下文为空或用户希望继续工作时，系统性地阅读状态文档、验证理解、并在继续工作前产出恢复报告。
resume 分为四个阶段：(A) 执行恢复链（身份锚定 → 必读 → 任务条件读）、(B) 产出恢复报告（7 节结构化综合）、(C) 理解验证（5 道强制问题证明理解）、(D) 恢复决策。
interactive（默认）：将恢复报告呈现给用户确认后再继续。auto：无需用户交互执行完整流程，应用自动恢复决策树（≤7 天新鲜 + 无边缘条件 = 继续；否则 = 等待用户）。
→ 详见 resume.md。
/doc-harness recall [query]
从项目 Doc Harness 文档层次结构中检索信息。在已注册文档中进行系统化搜索，返回结构化、带来源引用的结果。
查询类型：
- 状态/计划："认证当前计划是什么？" → 搜索 CURRENT_STATUS + headlights
- 历史/决策："我们为什么选 PostgreSQL？" → 搜索 WORKLOG + tire tracks
- 文件查找："找所有关于缓存的文档" → 搜索 FILE_INDEX
- 综合："所有关于认证的讨论" → 跨所有层综合搜索
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago Changed · +191 tokens per session 65e09174f134
- 10d ago First seen · 108 lines · 0 tokens per session scan A 82345fcbcc88
doc-harness is a skill published in the GitHub repository cilidinezy-commits/doc-harness (4 stars, last pushed 9d ago), licensed MIT. It adds 191 tokens to every session and 2,059 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
state
Use when the user says 'update state', 'project state', 'where was I', or at session start to load current context.
harness-step3-session-management
A project setup for keeping an agent’s work state in files between separate chat sessions. It creates a task list, a progress summary, and a script for checking the project environment.
project-context-loader
Load relevant memory, decisions and contacts for the current project at the start of a session.
sweep-now
Run the three read-mostly vault sweeps for stale tracker items, dormant-note candidates, and broken wikilinks, then append one dated result to SWEEPLOG.md.
operator-setup
The AIgent first-run operator-business interview. Learn the operator's business in three plain questions, write the first memory, and play back a briefing that reads personal, not templated. This is the "it remembers me" moment. A distinct skill that LAYERS ON TOP of the base /setup (which is left intact), it does not…
continue
Use when you need the Codex Copilot equivalent of Claude Copilot /continue: resume paused or previous work from tc tasks, work products, memory, and stream context.