메뉴
Create a new technology page in the technologies/ directory. Use when adding a new company, model, or product to the Lablab tech index — researches the technology, drafts the MDX, and places it in the correct location.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | create-tech-page |
| description | Create a new technology page in the technologies/ directory. Use when adding a new company, model, or product to the Lablab tech index — researches the technology, drafts the MDX, and places it in the correct location. |
| compatibility | Designed for Claude Code. Requires WebSearch or Exa MCP for research. |
Ask the user (if not already provided):
technologies/[provider-slug]/index.mdxtechnologies/[provider-slug]/[model-slug].mdx (provider dir may already exist)Before writing anything, check whether the page already exists:
Glob: technologies/[provider-slug]/**/*.mdx
Use this search order:
mcp__exa__*) — semantic search for official docs and overviewsmcp__brave__*) — keyword searchQueries to run:
"[technology name]" site:[official-domain] docs OR documentation"[technology name]" overview OR about OR features"[technology name]" GitHub"[technology name]" API OR SDKCollect:
Do not invent facts. If something is not found, omit it rather than guess.
File path rules:
technologies/[provider-slug]/index.mdxtechnologies/[provider-slug]/[model-slug].mdxllama-4, groq, claude-sonnet)Branch: Check the current git branch. If already on a suitable feature branch (e.g. add-[slug]), use it. Otherwise, remind the user to create one:
git checkout -b add-[slug]
index.mdx)---
title: "[Company Name]"
description: "[One sentence: what the company does and why it matters to AI builders. 20–40 words.]"
---
# [Company Name]
[2–3 sentence intro: founding story, core mission, and what makes it relevant to developers.]
| General | |
|---------|--|
| **Company** | [Company Name](https://company.com) |
| **Founded** | [Year] by [Founder(s)] |
| **Headquarters** | [City, Country] |
| **Website** | [company.com](https://company.com) |
| **Documentation** | [Docs title](https://docs.company.com) |
| **GitHub** | [github.com/org](https://github.com/org) |
| **Type** | [Category, e.g. "AI Infrastructure", "LLM Provider"] |
---
## Core Products
### [Product 1 Name]
[2–3 sentences describing what it does and who it is for.]
### [Product 2 Name]
[2–3 sentences.]
---
## Developer Resources
[1–2 sentences introducing the developer ecosystem.]
<TechTutorials/>
### Helpful Links
- [Documentation](https://docs.company.com) — official API and SDK reference
- [GitHub](https://github.com/org) — open-source repos and examples
- [Discord](https://discord.gg/...) — community support *(if available)*
- [API Console / Playground](https://console.company.com) — start building
---
## Key Features
**[Feature 1]**
[1–2 bullet points or sentence.]
**[Feature 2]**
[1–2 bullet points or sentence.]
---
## Use Cases
**[Use Case 1]**
[1–2 sentences.]
**[Use Case 2]**
[1–2 sentences.]
[model-slug].mdx)---
title: "[Model/Product Name]"
author: "[provider-slug]"
description: "[One sentence: what the model/product is, its architecture or key capability, and availability. 25–50 words.]"
---
# [Model/Product Name]
[2–3 sentence intro: what it is, what makes it technically notable, and where it fits in the ecosystem. Link the name to the official announcement or product page.]
| General | |
|---------|--|
| **Release date** | [Day Mon Year] |
| **Developer** | [Company Name] |
| **Type** | [e.g. "Open-weight multimodal LLM", "Text-to-speech API"] |
| **License** | [License name, or "Commercial API"] |
| **GitHub** | [org/repo](https://github.com/org/repo) *(if open-source)* |
| **Documentation** | [Docs title](https://docs.url) |
---
## Core Features
- **[Feature 1]** — [one-line explanation with source link if available].
- **[Feature 2]** — [one-line explanation].
- **[Feature 3]** — [one-line explanation].
---
## [Model Variants / Pricing Tiers] *(include only if the product has multiple variants)*
| Variant | [Key Param] | [Key Param] | Best for |
|---------|-------------|-------------|----------|
| **[Variant A]** | [value] | [value] | [use case] |
| **[Variant B]** | [value] | [value] | [use case] |
---
## Tools and Resources
- **[Resource 1]** — [what it is and link].
- **[Resource 2]** — [what it is and link].
- **[Quickstart / Playground]** — [link].
---
## Ecosystem and Integrations
- [Integration or deployment context 1].
- [Integration or deployment context 2].
---
[1–2 sentence closing: how to get started, with 1–2 direct links.]
Run these checks on the draft before writing the file:
title and description present; author present on model pages<TechTutorials/> included on company index pages (in the Developer Resources section)[insert X], no TBD unless genuinely not yet announcedWrite the drafted MDX to the correct path. Confirm the full path to the user before writing.
Report:
technologies/[path]publish-check if the page is ready to ship, or note what still needs filling inGenerate an article or tutorial concept from a source (tweet, article, or raw idea). Use when you have a piece of content or an idea and want to turn it into a lablab content concept — outputs a working title and brief ready to hand off to content-start.
Run the publishing checklist for a Lablab article or tutorial before marking it as Done. Use when a draft is complete and ready to be signed off — checks all sections are written, code blocks work, images use Cloudinary, and Notion task is updated.
Generate a tutorial concept directly from a hackathon page. Takes a hackathon URL or pasted page content, extracts partners, tracks, timeline, and challenge requirements, then identifies the strongest tutorial angle a participant could actually build and submit. Outputs a ready-to-write brief and Notion task. Hands off to content-start.
Build and verify a working code project before writing a tutorial. Use before drafting any tutorial that involves code — scaffolds the project, implements each component step by step, and confirms end-to-end functionality. Only hand off to writing once the code actually runs.
Run the mandatory pre-writing research protocol before starting any Lablab article or tutorial. Use when about to write or draft a new article or tutorial — checks for duplicates, gathers sources, confirms the angle, and sets the Notion task to In Progress.
Apply targeted SEO improvements to a finished article or tutorial without changing its tone or voice. Reads this week's keyword clusters from the Notion SEO drops database (generated by the weekly SEO agent), matches relevant clusters to the piece, then makes surgical changes only — frontmatter, one bridge paragraph, one internal link, and an optional FAQ for tutorials. Does not rewrite sections or stuff keywords throughout the body.