← 返回 Skills 市场
thierrypdamiba

DeepContent

作者 thierrypdamiba · GitHub ↗ · v2.0.0 · MIT-0
cross-platform ✓ 安全检测通过
177
总下载
0
收藏
0
当前安装
29
版本数
在 OpenClaw 中安装
/install deepcontentskill23223
功能描述
AI content marketing pipeline. Generate branded LinkedIn, X, and Reddit posts from any URL. Trigger on: "make a post from this", "turn this into content", "g...
使用说明 (SKILL.md)

DeepContent

Turn any URL into branded social media posts. The API does the heavy lifting. Your job is routing and UX.

API

Base: https://deepcontent-api.scaleintelligence.workers.dev OpenAPI spec: {BASE}/api/docs/openapi.json Auth: Authorization: Bearer {DEEPCONTENT_API_KEY}

Consult the OpenAPI spec for endpoint details, request bodies, and response shapes. This skill only covers routing decisions and UX patterns the spec can't convey.

Frontend

Base: https://deepcontent-frontend.scaleintelligence.workers.dev

Intent routing

User says Route to
URL or "make a post from this" Quick generate
"run a recipe" or wants multi-block output Full synthesis
"what should I write about" or "find topics" Topic discovery
"create a brand" or company URL with no existing brand Brand onboarding
"show my brands" Brand list
"show my posts" or "recent content" Post management
"invite [email]" or "add [person] to the team" Team invite
"how many credits" or "billing" Status
Lost or confused Help
No API key in session Device auth

UX rules

Auth

  • No key? POST /api/v1/connect/init to get a device code. Show: {FRONTEND}/connect/{code}
  • Poll GET /api/v1/connect/status/{code} until completed. Store the key for the session.
  • Get org ID from GET /api/v1/auth/me (returns orgId). Needed for invites.

Brand resolution

Every generate/topics flow needs a brand. Resolve it first:

  • Zero brands: guide to create one. Link: {FRONTEND}/dashboard/brands/onboarding
  • One brand: use it automatically, don't ask
  • Multiple brands: ask which one by name
  • User named one: match it

After generating content

  • Show the content first, not explanations
  • Each platform gets its own section header
  • ALWAYS link to: {FRONTEND}/dashboard/posts/{post_id} (path is always /dashboard/posts/, never /dashboard/linkedin/ or /dashboard/x/)
  • Show remaining credits. Link: {FRONTEND}/dashboard/billing
  • Ask: "Want to edit anything before approving?"

After creating a brand

  • Show: name, industry, audience, voice/tone
  • ALWAYS link to: {FRONTEND}/dashboard/brands/{id}
  • Ask which platforms (linkedin, x, reddit)
  • Brand starts as "draft". Confirm with POST /api/v1/brand-onboarding/{id}/confirm

After discovering topics

  • Show topics with title, relevance, source URL
  • Link each to: {FRONTEND}/dashboard/topics/{topic_id}
  • Ask: "Want me to generate content from any of these?"

Links (always include the relevant one)

Action Link
View/edit a post {FRONTEND}/dashboard/posts/{post_id}
Edit a brand {FRONTEND}/dashboard/brands/{id}
Create a brand {FRONTEND}/dashboard/brands/onboarding
View a topic {FRONTEND}/dashboard/topics/{topic_id}
All topics {FRONTEND}/dashboard/topics
All brands {FRONTEND}/dashboard/brands
Billing/credits {FRONTEND}/dashboard/billing
Team settings {FRONTEND}/dashboard/settings
Dashboard {FRONTEND}/dashboard
Sign up {FRONTEND}/sign-up

Slow endpoints

These take longer than usual. Use --max-time accordingly:

  • POST /api/v1/brand-onboarding/generate: 30-60s (scrape + AI analysis). Use 120s timeout.
  • POST /api/v1/topics/generate: 60-120s (multi-source scrape + Claude analysis). Use 180s timeout.
  • Brand onboarding: check for existing brand first (GET /api/v1/brands). Ask user to update or create new.

Learning

  • After edits, note what the user changed. Store as memory for next time.
  • If they consistently make the same edit (e.g. always remove closing questions), suggest updating the brand voice via PATCH /api/v1/brands/{id}.
  • Track: preferred platform, default brand, edit patterns. Skip questions when preferences are clear.

Response style

  • Content first, not explanations
  • Platform name as section header
  • 300 char preview max in Discord
  • End with dashboard link + credits
  • No celebrations, no exclamation marks
安全使用建议
This skill is coherent with its stated purpose, but before installing you should: (1) Verify the DEEPCONTENT_API_KEY scope and create a least-privilege key if possible. (2) Ask where the agent will store the session key and any persistent 'memory' (encrypted agent vault, ephemeral session only, or long-term memory?) and whether you can opt out of memory. (3) Test with non-sensitive URLs first — anything you send (including scraped page content) may be stored by the skill or the service. (4) Verify the frontend/api domains are legitimate and review the OpenAPI spec at the provided base URL to confirm endpoints and timeouts. (5) Confirm retention and deletion policies for generated posts, stored edits, and any org/team invites. If these answers are unsatisfactory, do not provide a production API key or enable persistent memory.
功能分析
Type: OpenClaw Skill Name: deepcontentskill23223 Version: 2.0.0 The skill is a standard integration for an AI content marketing service hosted on Cloudflare Workers (scaleintelligence.workers.dev). It uses Bash to interact with a documented API for generating social media posts and managing brand identities. The instructions in SKILL.md are well-structured, focusing on UX routing, authentication flows, and API timeouts without any signs of malicious intent, data exfiltration, or prompt injection attacks.
能力评估
Purpose & Capability
Name/description (content marketing pipeline -> generate social posts) align with what the SKILL.md asks the agent to do: call the DeepContent API, route intents, and present UX. The only required credential (DEEPCONTENT_API_KEY) is appropriate for an API-backed service.
Instruction Scope
Instructions are focused on routing/UX and API calls. However, the docs instruct the agent to 'Store the key for the session' and to 'note what the user changed' and store edits as memory for future runs — this implies persistent storage of user data and the API key. There are no directives to read unrelated files or environment variables, nor to transmit data to endpoints outside the stated API/frontend.
Install Mechanism
Instruction-only skill with no install spec and no code files. No binaries or downloads are requested, so nothing is written to disk by the skill itself.
Credentials
Only a single API key (DEEPCONTENT_API_KEY) is declared and used as the primary credential. That is proportionate and expected for the skill's purpose; no unrelated secrets or config paths are requested.
Persistence & Privilege
The skill does not request always:true and is user-invocable. But it explicitly tells the agent to store the API key in-session and to persist user edits/preferences as memory — this increases persistence and potential long-term access to user data. Confirm what storage is used and whether storing the API key is encrypted/consented.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install deepcontentskill23223
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /deepcontentskill23223 触发
  4. 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
v2.0.0
Version 2.0.0 of deepcontentskill23223 introduces a major skill doc overhaul for clarity and precision: - Docs rewritten for concise routing logic and UI/UX guidance; detailed endpoint-by-endpoint flows removed. - Clearer intent routing table: maps user intent to action for faster processing. - Enhanced UX rules: strict linking patterns, content-first responses, and redundancy avoidance on brand/platform selection. - Brand, billing, and team management streamlined into always-present, standardized link formats. - Explicit slow-endpoint handling guidance. - Consistency enforced for response style, link usage, and interface prompts.
v1.6.1
- Updated post link instructions: For generated posts, always use the path /dashboard/posts/{post_id}, regardless of platform. - Documentation and flow clarifications to ensure correct post linking in the quick generate process.
v1.6.0
deepcontentskill23223 v1.6.0 - Updated team member invitation flow: now requires fetching the org ID from `GET /api/v1/auth/me` before sending invites. - Corrected manage posts dashboard links to be specific per platform (LinkedIn, Reddit, X). - No other user-facing changes.
v1.5.0
- Updated version to 1.5.0. - Changed the post-generation flow: dashboard link for generated posts now directs to `/dashboard/posts/{post_id}` instead of platform-specific pages. - No other functional or documentation changes detected.
v1.4.3
deepcontentskill23223 v1.4.3 - Updated post generation flow: platform post links in "Quick generate" now direct to `{FRONTEND_URL}/dashboard/{platform}` (e.g., `/dashboard/linkedin`) instead of linking by `post_id`. - Version bump to 1.4.3 in documentation.
v1.4.2
Version 1.4.2 - Improved the "Discover topics" section: now topic SSE events include a topic_id and direct links to each topic (`/dashboard/topics/{topic_id}`). - Clarified navigation by adding a link to view all topics in the dashboard. - No changes to functional endpoints or core flows. Documentation update only.
v1.4.1
Version 1.4.1 - Added timing and performance guidance to the "Discover topics" flow (endpoint can take 60-120 seconds; recommends using `--max-time 180` with curl). - No other major changes; documentation and flows remain essentially the same.
v1.4.0
Version 1.4.0 - Added a new "Learning from usage" section describing how to observe user content edits and store preferences for brand-specific future outputs. - Outlined steps to build a mental model of individual user habits and apply these to skip repetitive questions and streamline generation. - No functional changes to endpoints or workflows; core flow and response guidelines remain the same.
v1.3.3
- Added note clarifying that the brand creation API endpoint can take 30–60 seconds, with suggestion to use `--max-time 120` in curl calls. - No code or functional changes; documentation update only. - Version bump to 1.3.3.
v1.3.2
- Updated brand creation flow: responses must now always include a direct link to the just-created brand’s dashboard page. - No other changes; all existing flows and guidance are unchanged.
v1.3.1
- Improved brand selection logic in quick generate and topic discovery: now automatically uses a single brand, asks if multiple, or matches by name if specified. - Updated docs for clearer instructions on handling brand choice in content and topic generation flows.
v1.3.0
Version 1.3.0 expands content discovery and management features. - Added triggers and flows for "discover topics", "show my posts", and "what should I write about". - Introduced topic discovery via `/api/v1/topics/generate`, enabling content idea generation based on brand. - Added support for managing (listing, editing, deleting) generated posts across platforms. - Updated quick generate and brand creation flows for more precise options (e.g. subreddit targeting on Reddit, X thread format, draft brand status). - New billing and plan flow for user status and upgrades. - Expanded help and usage examples.
v1.2.0
- Adds duplicate brand detection when creating a brand: skill now checks for existing brands with similar name or URL before proceeding. - If a similar brand exists, users can choose to update the existing brand or create a new one. - Updates "Create a brand" flow to include an option for refreshing an existing brand via a new refill endpoint.
v1.1.0
- Added "Quick generate" flow for single-platform content generation from a URL. - Introduced "Full synthesis" flow for multi-block, recipe-based content outputs. - Updated documentation to clarify when to use quick generate vs. full synthesis. - No changes to configuration, triggers, or basic feature set. - Minor structural and documentation improvements.
v1.0.1
- Updated post generation flow: each platform's generated post now links directly to its unique history page using `post_id`. - Clarified that SSE `done` events in content generation include `post_id` for improved tracking. - No changes to API endpoints or skill triggers; documentation update only.
v1.0.0
Major update: Adds team invites, URL scraping, and improved help guidance. - Added team member invite flow and dashboard/team management link. - New flow to scrape and summarize content from any URL. - Enhanced help and usage instructions for common user actions. - No changes to API config, content flows, or response style.
v0.9.1
Version 0.9.0 introduces an updated version number in documentation. - Updated SKILL.md version from 0.6.0 to 0.9.0. - No other content or functional changes.
v0.9.0
- Major refactor: streamlined to a single SKILL.md file, removing 46 supporting files and legacy templates. - Configuration and usage flows in SKILL.md are clarified and expanded, with improved steps for device auth, brand onboarding, and credit management. - Brand creation and management flow updated to use /brand-onboarding/generate and detailed summary output. - Response and link formats revised for more direct user guidance and dashboard navigation. - Now mandates DEEPCONTENT_API_KEY via env, reflecting new requirements in metadata.
v0.5.3
Docs: correct brand creation flow (use POST /api/v1/brands with inputs+graph); document known API bugs (brands 500, generate SSE DB error, onboarding endpoint hangs); add workaround notes
v0.8.0
- Expanded "Create a brand" flow: the onboarding response now includes a summary of brand identity (name, industry, description, audience, voice), and the summary is shown to the user. - Updated instructions to show more detailed brand info and edit link after brand creation. - No changes to code or API flows; SKILL.md documentation only.
元数据
Slug deepcontentskill23223
版本 2.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 29
常见问题

DeepContent 是什么?

AI content marketing pipeline. Generate branded LinkedIn, X, and Reddit posts from any URL. Trigger on: "make a post from this", "turn this into content", "g... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 177 次。

如何安装 DeepContent?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install deepcontentskill23223」即可一键安装,无需额外配置。

DeepContent 是免费的吗?

是的,DeepContent 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

DeepContent 支持哪些平台?

DeepContent 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 DeepContent?

由 thierrypdamiba(@thierrypdamiba)开发并维护,当前版本 v2.0.0。

💬 留言讨论