Skills Plugins MCP Prompt Model 博客 我的中心
生活与工具 #javascript #browser #automation #ai

obscura

Operate and validate Obscura for JavaScript page loading, stealth browsing, anti-fingerprinting, tracker blocking, screenshots and visual comparison, CDP automation with Puppeteer or Playwright, screencasting, PDF export, MCP browser interaction, and web extraction. Use when running Obscura against deterministic fixtures or real sites, diagnosing rendering, geometry, resource, identity, or transport failures, or choosing the correct CLI, CDP, MCP, rendering, or stealth workflow.

DeepseekModel 官方收录技能 质量 优秀 · 90 v1.0.0

获取

https://deepseekmodel.com/api/download.php?id=h4ckf0r0day-obscura-skills-obscura-skill-md&format=skill
下载 .skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用
.skill 文件中 system_prompt 字段的实际内容。
name obscura description Operate and validate Obscura for JavaScript page loading, stealth browsing, anti-fingerprinting, tracker blocking, screenshots and visual comparison, CDP automation with Puppeteer or Playwright, screencasting, PDF export, MCP browser interaction, and web extraction. Use when running Obscura against deterministic fixtures or real sites, diagnosing rendering, geometry, resource, identity, or transport failures, or choosing the correct CLI, CDP, MCP, rendering, or stealth workflow. Obscura Use Obscura as a lightweight, stealth-capable Rust headless browser for automation. It embeds V8, owns the DOM and rendering pipeline, and exposes Chrome DevTools Protocol workflows without launching Chromium. Treat rendering and stealth as first-class, complementary capabilities. Build variants Official release archives and Docker images include rendering. For a source checkout, build release mode with the render feature: CARGO_INCREMENTAL=0 CARGO_BUILD_JOBS=2 cargo build --release -p obscura-cli --bins --features render Build rendering and stealth together for the wreq/BoringSSL transport, browser-identity protections, and tracker blocking: CARGO_INCREMENTAL=0 CARGO_BUILD_JOBS=2 cargo build --release -p obscura-cli --bins --features render,stealth Build without rendering when only DOM, extraction, or CDP automation is needed: CARGO_INCREMENTAL=0 CARGO_BUILD_JOBS=2 cargo build --release -p obscura-cli --bins --no-default-features CARGO_INCREMENTAL=0 CARGO_BUILD_JOBS=2 cargo build --release -p obscura-cli --bins --no-default-features --features stealth Use ./target/release/obscura in the commands below when working from source. Use stealth The stealth build keeps the complete rendering, screenshot, screencast, PDF, CDP, and MCP surface. It adds a consistent browser fingerprint across TLS, HTTP headers, user agent, navigator, and WebGL surfaces; masks navigator.webdriver ; masks patched native functions; and blocks the built-in tracker-domain list. Enable stealth at runtime with the global --stealth flag. It applies to fetch , serve , scrape , and mcp , before or after the subcommand: obscura --stealth fetch https://example.com --screenshot page.png obscura serve --stealth --port 9222 The runtime flag needs a render,stealth build for the wreq/BoringSSL transport. Fetch, evaluate, and capture obscura fetch https://example.com --dump text obscura fetch https://example.com -- eval "document.title" obscura fetch https://example.com --screenshot page.png obscura fetch https://example.com \ -- eval "window.scrollTo(0, document.documentElement.scrollHeight)" \ --screenshot bottom.png The CLI screenshot path writes PNG and accepts one URL. An omitted --wait uses adaptive settling with a five-second cap. An explicit --wait N is a fixed delay of N seconds. --timeout is also measured in seconds and bounds navigation separately. Use --dump original for a binary or raw HTTP response that should bypass DOM and JavaScript processing. Use scrape for many URLs when the requested output does not require one screenshot per URL. Drive CDP Start the server: obscura serve --port 9222 Connect Puppeteer with puppeteer-core or Playwright with chromium.connectOverCDP . Standard page.screenshot() supports viewport and full-page capture; page.pdf() produces raster-backed print output. Scroll with page JavaScript before a viewport capture when the user wants a lower section of the page. Use raw CDP for Page.captureScreenshot , Page.startScreencast , and Page.printToPDF . A screencast client must acknowledge every Page.screencastFrame using Page.screencastFrameAck . Frames are driven by page activity; this is not fixed-frame-rate desktop capture. PDF output supports paper dimensions, margins, landscape, scale, backgrounds, and page ranges. It does not currently provide selectable text, tagged PDF, outlines, headers/footers, or complete CSS paged-media behavior. Drive MCP Run obscura mcp for stdio or obscura mcp --http --port 3000 for HTTP. Navigate first, then inspect or interact with the current page. Refresh a snapshot or interactive-element listing after navigation, clicking, scrolling, or a framework rerender because element references may have changed. Render-enabled MCP builds expose browser_screenshot as an MCP PNG image and browser_pdf as an embedded PDF resource. MCP does not stream screencast frames; choose CDP for that workflow. Validate visual behavior Start with a deterministic fixture that isolates the behavior. Then test a broad real-site set at both initial and scrolled positions. For an engine comparison, keep viewport, device scale, user agent, network inputs, settle policy, scroll, animation time, and capture boundary identical. Confirm both engines navigated successfully and produced nonblank images before interpreting a diff. Treat a pixel metric as a regression tripwire, not a verdict. Inspect resource completion, box geometry, line wrapping, structural edges, clipping, and fixed or sticky behavior, then reduce real failures to a fixture. Do not introduce hostname-specific rendering logic. Set expectations accurately Obscura supports many common layout and paint paths but is not a bundled Chrome build. Long-tail CSS, service workers, some Web APIs, native media, GPU or compositor effects, PDF structure, and platform font rasterization can differ from Chromium. Preserve the project's existing positioning and published benchmark claims when editing its documentation. Use the benchmark suite and matched inputs when adding or updating performance or fidelity measurements.
Agent 识别该技能的关键词,点击任意一个即可复制。

该技能未提供触发词。

下载的 .skill 包内含以下字段。
字段 说明
format格式标识(skill/v1)
skill_id技能唯一 ID
name技能名称
version版本号
description技能描述
category所属分类(数组)
trigger_words触发词列表
tags标签列表
source来源标识
source_url来源链接(本页地址)
exported_at导出时间(每次下载生成)
system_prompt系统提示词正文
model_config模型参数:provider / model / temperature / max_tokens / top_p
examples示例
install_guide各平台导入说明(Coze / Dify / Claude / 自定义框架)
同一份技能可按不同平台格式导出。
.skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用 下载
.skillpro 增强格式,额外含脚本 / 工具 / 依赖 / 钩子占位 下载
.json 纯 JSON 导出,只含 system_prompt 与模型参数 下载
Coze 带 frontmatter 的 Markdown,Coze 平台导入用 下载
Dify Dify DSL,创建应用后直接导入 下载

每日精选 Skill 推荐,免费送到你邮箱

输入邮箱,每天接收一个精选 AI Agent 技能推荐。完全免费,持续更新。

验证码 --

提交后我们会发送一封确认邮件,点击邮件里的链接才会开始收信。

完全免费,取消任意时间。我们不会发送垃圾邮件。