Token导航 LogoToken导航TokenDH.com
研究检索操作浏览器github未标认证来源可访问许可证需确认审计异常

superwall-editorsuperwall editor 搜索

Agent Skill

superwall-editor 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

1,717

周安装

73

GitHub Stars

17

下载量

602
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:superwall-editor(superwall editor 搜索)
来源仓库:https://github.com/superwall/skills
仓库路径:skills/superwall-editor
安装命令:
npx skills add https://github.com/superwall/skills --skill superwall-editor
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/superwall/skills --skill superwall-editor

简介

superwall-editor 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。

  • 适用于需要根据关键词或任务场景进行信息检索的场景,如内容编辑或资料整理。
  • 通过关键词输入和来源仓库筛选,Agent 可返回结构化候选结果供进一步处理。
  • 安装前需确认权限范围和维护状态,注意是否触发联网或文件读写操作。
  • 建议结合原始 README 核验具体用法,避免直接使用未经验证的输出结论。

SKILL.md

Superwall Paywall Editor

Paywalls are built in a browser editor that exposes its tools over an authenticated relay. This skill drives that relay from the CLI — the exact same surface the MCP gateway uses — so every tool you invoke runs inside the live browser session the user has open.

When to use

  • The user wants to build, edit, or review a Superwall paywall, onboarding, or web2app flow.
  • The user pastes a pairing code and asks you to take over editing.
  • The user asks "what tools can you run right now?" — discover them via the browser, not from memory.

Start here: attach, then discover

Never assume a tool name or signature from memory. The browser is the source of truth and its tool set changes across releases.

Preferred API launch flow:

  1. Create an auto-expose URL: scripts/sw-editor.sh expose --application-id <id> --paywall-id <id> --agent-name <agent> --open --wait
  2. Ask the user to complete browser authorization if prompted. The editor auto-exposes; do not ask them to click the expose button.
  3. Discover what is available right now: scripts/sw-editor.sh tools
  4. Invoke tools: scripts/sw-editor.sh call <tool-name> --args '<json>'

Fallback manual flow:

  1. Ask the user for the pairing code shown in the editor UI.
  2. Attach: scripts/sw-editor.sh attach <pairing-code>
  3. Continue with tools and call.

Full CLI reference: references/cli.md.

How to build and edit

Orchestration rules

  • Always establish an attachment before editing. Use expose --open --wait when possible, otherwise use attach <pairing-code>. tools, call, status, release all require an attached session.
  • Prefer expose --open --wait when you have SUPERWALL_API_KEY, an application id, and a paywall id. It uses the same relay as manual pairing but removes the human pairing-code step.
  • Before calling a tool you have not used this session, run tools to confirm it exists and to read the current parameter schema. Tools are defined in the browser bundle — an updated editor ships new or renamed tools without any change to this skill.
  • Use get_screenshot (if present in the tool list) every 2–3 modifications to verify. Don't fly blind.
  • Prefer semantic tools (update_styles, set_text_content, set_dynamic_value, move_nodes) over re-running write_html on existing structure. See references/workflow.md.
  • Prefer native sw-* elements over hand-rolled <div> recreations whenever the UI represents a semantic control. See references/native-elements.md.
  • When parsing CLI output, use jq — never Python. Example: sw-editor.sh call get_subtree --args '...' | jq -r '.content[0].text'
  • Release when the user is done: scripts/sw-editor.sh release.

When things go wrong

  • session_not_ready: the browser disconnected or reloaded. Ask the user to bring the editor tab back, then re-attach (the pairing code rotates — they'll need to read you the new one).
  • session_locked: another client is already attached. The user either attached from another MCP client, or a previous CLI attachment wasn't released. They can detach from the editor UI and you can retry.
  • unauthorized: the controller token is stale — re-attach with a fresh pairing code.
  • attach_failed: provide a valid current pairingCode: pairing codes expire after ~10 minutes and rotate on detach. Ask the user to show you the current one.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.14%
按下载量换算212

Claude

30.8%
按下载量换算185

Cursor

16.57%
按下载量换算100

Gemini CLI

7.87%
按下载量换算47

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

未通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills