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 agentmods add skills/clawcap/manobrowser/chrome-workflow-buildnpx skills add ClawCap/ManoBrowser --skill chrome-workflow-buildgit clone --depth 1 https://github.com/ClawCap/ManoBrowserWrote 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/clawcap/manobrowser/chrome-workflow-build)<a href="https://agentmods.dev/skills/clawcap/manobrowser/chrome-workflow-build"><img src="https://agentmods.dev/badge/skills/clawcap/manobrowser/chrome-workflow-build.svg" alt="Measured on agentmods" 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 | $0.00134 | $0.03925 |
| Opus 5 | $0.00067 | $0.01962 |
| Sonnet 5 | $0.00027 | $0.00785 |
| Haiku 4.5 | $0.00013 | $0.00392 |
Grade A, and why
chrome-workflow-build 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 4d 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 — 323 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Chrome工作流构建器
概述
这是一个元技能(Meta-skill),支持双模式将浏览器自动化操作转换为可复用的Claude Skills:
- 模式1(执行并记录):通过智能记录、分析和生成,将实时执行的浏览器操作链路转换为Skill
- 模式2(DATASAVER录制转换):将DATASAVER录制工作流.json文件分析、验证并转换为Skill
规则1:严格按阶段顺序执行
- 模式1:阶段1(LOG.md) → 阶段2(WORKFLOW.md) → 阶段3(CREATOR.md)
- 模式2:阶段1(DATASAVER录制工作流转换模块) → 阶段2(CREATOR.md)
- 禁止跳过任何阶段,每个阶段必须完整执行后才能进入下一阶段
- 禁止偷懒:不能只做部分工作就声称完成
规则2:中间产物必须生成
每个阶段必须生成对应的输出文件,缺少任何一个都视为阶段未完成:
| 阶段 | 必须生成的文件 |
|---|---|
| 模式1-阶段1 | workflow-logs/[skill-name]-log.json |
| 模式1-阶段2 | workflow-workflows/[skill-name]-workflow.json |
| 模式2-阶段1 | workflow-workflows/[skill-name]-workflow.json |
| 最终阶段 | workflow-skills/[skill-name]/SKILL.md |
快速开始
本Skill支持双输入模式:
模式1:执行并记录(适用于新需求)
- 立即使用chrome-server MCP 提供的工具执行任务
- 将执行步骤记录到结构化日志并抽取的Workflow
- 基于Workflow生成可供未来使用的可复用Skill
模式2:DATASAVER录制工作流转换(适用于已有DATASAVER录制文件)
- 提供DATASAVER录制工作流.json文件
- 分析并验证转换为workflow.json
- 基于Workflow生成可供未来使用的可复用Skill
工作流程
模式1:执行并记录流程
阶段1:执行并记录
执行浏览器任务,同时记录每个工具调用和决策。 加载 LOG.md 并严格遵循它的指令工作
阶段2:提取工作流(仅在阶段1完成后)
分析执行日志以提取可复用的工作流模式。 加载 WORKFLOW.md 并严格遵循它的指令工作
阶段3:生成Skill(仅在阶段2完成后)
从提取的工作流创建、验证并生成新的Skill。 加载 CREATOR.md 并严格遵循它的指令工作
模式2:DATASAVER录制工作流转换流程
阶段1:DATASAVER录制工作流分析提取workflow
读取DATASAVER录制工作流.json文件,分析、验证并提取为workflow.json。 加载 DATASAVER.md 并严格遵循它的指令工作
阶段2:生成Skill
从生成的工作流创建、验证并生成新的Skill。 加载 CREATOR.md 并严格遵循它的指令工作
决策树
用户请求分析
├─ 1. 判断输入类型
│ ├─ 提供DATASAVER录制工作流.json文件? → 模式2: DATASAVER录制工作流转换流程
│ ├─ 描述浏览器操作需求? → 模式1: 执行并记录流程
│ └─ 其他 → 不使用此Skill
│
├─ 2. 模式1: 执行并记录流程
│ │
│ 阶段1:执行并记录(LOG.md)
│ ↓
│ 【确认点】展示最终执行结果,询问用户是否正确
│ ├─ 正确 → 进入Skill发现检查
│ └─ 有问题 → 根据用户反馈调整后重新执行
│ ↓
│ Skill发现检查(LOG模式-执行后立即发现)
│ 加载 SKILL_DISCOVERY.md 执行智能发现
│ ├─ 发现高度匹配Skill(>=90分)→ 强制询问用户 → 等待用户决策
│ │ ├─ 用户确认"满足需求" → 停止创建流程 + 文字指导使用 → 结束(不保存日志)
│ │ └─ 用户选择"不满足" → 保存日志 → 继续阶段2和3
│ └─ 无匹配(<90分)→ 保存日志 → 直接继续阶段2和3(不提示)
│ ↓
│ 阶段2:提取工作流(WORKFLOW.md)
│ ↓(自动继续)
│ 阶段3:生成Skill(CREATOR.md)
│ ↓
│ 完成(包括验证和发布报告,用户确认后保存文件并发布服务端(如果存在skill发布tool))
│
└─ 3. 模式2: DATASAVER录制工作流转换流程
│
阶段1:DATASAVER录制工作流分析转换(DATASAVER录制工作流转换模块)
- 读取DATASAVER录制工作流.json文件
- 提取description和业务需求
↓
Skill发现检查(DATASAVER模式-读取后立即发现)
加载 SKILL_DISCOVERY.md 执行智能发现
├─ 发现高度匹配Skill(>=90分)→ 强制询问用户 → 等待用户决策
│ ├─ 用户确认"满足需求" → 停止创建流程 + 文字指导使用 → 结束
│ └─ 用户选择"不满足" → 继续转换流程
└─ 无匹配(<90分)→ 直接继续转换流程(不提示)
↓
- 分析并抽取必要步骤
- 逐步执行验证
- 生成workflow.json
↓
阶段2:生成Skill(CREATOR.md)
↓
完成(包括验证和发布报告,用户确认后保存文件并发布服务端(如果存在skill发布tool))
What ships with it
19 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.
- assets/datasaver_workflow_template.json 6.6 KB
- assets/log_template.json 10 KB
- assets/skill_template.md 6.9 KB
- assets/workflow_template.json 6.0 KB
- CREATOR.md 13 KB
- DATASAVER.md 26 KB
- LOG.md 18 KB
- references/CONTEXT_ISOLATION.md 6.7 KB
- references/CSS_SELECTOR.md 7.4 KB
- references/DATASAVER_CONVERSION.md 19 KB
- references/DATE_PICKER_HANDLING.md 15 KB
- references/EXECUTOR_RULES.md 14 KB
- references/FILTER_STEP.md 11 KB
- references/PUBLISH_SKILL.md 3.8 KB
- scripts/init_skill.py 11 KB runs code
- scripts/validate_log.py 18 KB runs code
- scripts/validate_skill.py 7.2 KB runs code
- SKILL_DISCOVERY.md 3.1 KB
- WORKFLOW.md 18 KB
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.
- 4d ago First seen · 323 lines · 134 tokens per session scan A 4ba6b9cb8aba
chrome-workflow-build is a skill published in the GitHub repository ClawCap/ManoBrowser (9 stars, last pushed 4mo ago), licensed MIT. It adds 134 tokens to every session and 3,925 once invoked, about $0.0007 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
aipex-browser
AI-powered browser automation using the AIPex Chrome Extension via MCP bridge. Use this skill when the agent needs to control a Chrome browser — navigating pages, clicking elements, filling forms, capturing screenshots, managing tabs, or downloading content — by connecting to the AIPex MCP bridge.
just-scrape
Search, scrape, crawl, extract structured data, and monitor web pages via the ScrapeGraph AI CLI. Use when the user asks to search the web, scrape a webpage, grab content from a URL, extract JSON from a site, crawl documentation or site sections, monitor a page for changes, inspect request history, check ScrapeGraph…
browser-relay
Control the Chrome the user already has open and logged in through the Browser Relay CLI, without launching a separate automation browser or taking over the foreground tab. Use when an agent needs to work with existing sessions, cookies, extensions, SSO or intranet pages, or a browser on another machine. Prefer the…
owb
Open Web Bridge (OWB) — drive the user's own real browser with the owb command. Read pages behind their existing logins, gather and cross-check information, fill forms, walk multi-step flows, debug their site, audit responsive/accessibility behavior, and capture or reverse-engineer network traffic. Use this whenever…
website-explorer
Discover website capabilities from user behaviors. Learn APIs and automate workflows.
codex-browser
Control Chrome via Codex Desktop's browser bridge. 52 MCP tools for tabs, navigation, DOM, input, CDP, network, file upload, dialog handling, and diagnostics.