Growth Lab skill
A skills set for Claude Code and Codex: the agent studies demand, builds SEO pages and articles and reviews results
- 1,971 ★ on the repository
- Last commit
- Apache-2.0
Broad access
Why this level
- Service keys and WeChat credentials sit in local settings, and the agent works with them from the working folder
- Can create pages, submit URLs to a search engine and sync drafts to WeChat
- Collectors for third-party platforms drive a browser and can lead to limits or an account ban
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/growth-labDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Service keys and WeChat credentials sit in local settings, and the agent works with them from the working folder
- Can create pages, submit URLs to a search engine and sync drafts to WeChat
- Collectors for third-party platforms drive a browser and can lead to limits or an account ban
This is third-party code. Review the repository files before installing.
What it does
Growth Lab is a repository of skills that you open as a working folder in Claude Code or Codex. The agent session runs the process, skills hold the methods and files in the folder serve as long-term memory. You state the task in plain language, for example: understand this product and run its first growth loop. The skills sit in three layers. Collectors study the product, search demand and Bing Webmaster statistics. Executors create and review SEO pages, generate images, submit URLs to IndexNow and take screenshots. The third layer is the growth loops themselves, built as observe, act, review: each loop has its own memory folder that collects measurements, conclusions and next steps. Initial setup is handled by the onboard-growth-lab skill. It checks keys and external clients, shows a readiness table and asks what to enable.
Who it is for. For founders and marketers who run a product with Claude Code or Codex and want demand research, pages and result reviews in one working folder.
Good fit when
- You need to find the queries people use for your problem and publish informative SEO pages with a review step before release
- You need to submit new URLs to IndexNow and then compare impressions and clicks from search data
- You need to run Xiaohongshu notes or WeChat articles from collecting samples to the pre-publication check
- You want images for the materials and screenshots of finished pages in the same process
Not a fit when
- Your audience is on Yandex and Russian social networks: the ready collectors and publishers target Bing, Xiaohongshu, WeChat and Chinese platforms
- You need a permanent background service: the loop runs in a conversation with the agent and there is no separate server program
- You need one narrow task such as a single audit: a plain SEO skill without memory and loops is simpler
Example request
Check what Growth Lab is missing, set up SEO only and run the SEO page loop for my product: find queries, pick one and prepare a pageLimitations
- It works with Claude Code or Codex as the runtime. You clone the folder and open it in the agent; there is no package to install.
- Most capabilities need their own keys and credentials; the fields are listed in the .env.example and CONFIGURATION.md files.
- The Xiaohongshu collector works through a local browser MCP, and collectors for other platforms use an external MediaCrawler. Automated collection can trigger platform limits or an account ban, so it needs a separate spare account and Chrome 144 or newer.
- The author's results come from one run and depend on the product, demand and site authority.
- The main documentation is written in Chinese; an English version and a Russian translation are in the repository.
- The license is Apache-2.0.
How to disable. Close the folder in the agent and delete the repository clone. Loop memory sits in the memory folder, and secrets live outside the repository in .env.local or the system keychain; delete them separately if you created any.
FAQ
How do I install Growth Lab in Claude Code or Codex?
Clone the repository with `git clone https://github.com/tsingyuai/growth-lab.git`, open the folder in Claude Code or Codex and ask what Growth Lab can do. The onboard-growth-lab skill checks keys and suggests what to connect.
Which growth loops does Growth Lab include?
Three: SEO pages, Xiaohongshu notes and WeChat articles. The onboard-growth-lab readiness table shows which of them are ready to run with your keys.
Does Growth Lab publish articles by itself?
Only after an explicit opt-in. Xiaohongshu posts are published by a person. A WeChat article stays a draft by default: publishing needs an environment switch, approval in the wechat.yml file and the `--confirm-publish` flag in the command.
Where are data and agent memory stored?
In your working folder. Product data, collected samples, results and loop memory sit in the memory folder and nearby, and the project keeps no copy on its side. Secrets, cookies and browser profiles never enter memory.
Which keys does Growth Lab need?
It depends on the capabilities. SEO needs a Bing Webmaster key and an IndexNow key, images need an OpenAI or Gemini key, WeChat needs an app ID and secret. You decide what to enable, and anything unneeded can be bypassed for the session.
Discussion
No questions or reviews yet.
Related
SkillsMarketing skills for coding agents: SEO, content, email campaigns, ads, analytics and publishing through service APIs
AI Marketing Suite
AI Marketing Suite for Claude Code
A set of marketing skills for Claude Code: site audit, copywriting, emails, content calendars, competitor analysis and client reports
A library of 64 skills for founders: GTM, growth, SEO, email, graphics and video, installed with one npx command for Claude Code, Codex and more
Three experimental skills: adapt an article for channels with a source map, sort exported AI answers for GEO and spot-check hreflang
SEO article writer skill with images
SEO Article Writer
A 10-stage Claude Code pipeline: Wordstat semantics, top-5 competitor analysis, AI cover and illustrations, publishing to WordPress, Dzen and Prisma Blog
A Claude Code skill for the full Yandex-focused SEO content cycle, from keywords to AEO, with a Python engine and real Wordstat data
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Growth Lab (this entry) | no | 1,971 | ||
| Marketing Skills for AI agents | no | 57,199 | about fifty separate skills for CRO, copy and ads, with no loop memory or data collectors | |
| OpenSEO | yes | API key | 25,649 | an MCP server and skills for SEO research and audits, with no publish and review loops |
README in short
The idea behind the project: AI growth tools usually solve one local task and lose the product context. Growth Lab runs the full cycle from understanding the product to reviewing results, and the repository has sample prompts, setup through a conversation with the agent and a table of ready capabilities.
README badge
Original SKILL.md
--- name: run-seo-page-loop description: Run an SEO page observation-action-review loop with persistent Memory by coordinating demand research, page creation, adversarial review, image generation, IndexNow submission, and performance review. Use when taking an SEO page from opportunity discovery through publication, measurement, iteration, or continuing a previous SEO loop from Memory. --- # Run the SEO page loop Coordinate the loop inside the product workspace. Let the current Codex or Claude Code session control the work. Use `memory/run-seo-page-loop/` as this Model's persistent Memory. ```text Read Memory → Observe → Decide → Act → Review → Write Memory → Next observation ``` Read [memory.md](references/memory.md) before starting. Recover relevant observations, actions, outcomes, conclusions, and next-action recommendations. ## Boundaries - Keep this Model focused on when and why the loop moves between observation, decision, action, and review. - Delegate data-collection methods and source-specific interpretation to Collectors. - Delegate creation, implementation, publishing, inspection, and performance-review techniques to Executors. - Use Runtime-native browser, search, page inspection, screenshot, and local web-testing capabilities directly. - Add a Client only for an external API action the Runtime cannot perform natively. - Create no fixed schema, database, dashboard, workflow state, or task queue. - Store dated operational evidence, analysis, outcomes, and next-action recommendations in Memory. - Apply improvements to the loop itself directly to this Model. Keep methodology-change suggestions out of Memory. ## 1. Read Memory Read recent Memory entries and older entries relevant to the product, page, query family, or pending action. Establish what is already known, what was attempted, what happened, and which recommendation should now be tested. ## 2. Observe Invoke `$research-seo-demand` to collect and interpret current search demand and live SERP evidence. Combine it with product context and relevant Memory. When the loop begins from an existing page, invoke `$review-seo-performance` first to observe its current outcome. Persist useful raw evidence and a dated observation in `memory/run-seo-page-loop/`. ## 3. Decide Before choosing a page action, confirm that the current observation contains a competitor-page breakdown for every candidate query being considered. The breakdown must cover three to five relevant leading pages and include: - each page's search presentation and winning page shape; - a top-to-bottom description of its visible blocks; - reading and conversion hooks, information density, user value, and tone; - evidence, unique information, authorship, and negative quality signals; - an information-gain gap synthesized across the leading pages.