Analysis updated 2026-05-18
Generate a personal portfolio or job-search website from a LinkedIn profile using Codex.
Turn a set of approved reference images into a pixel-faithful implemented website.
Run the included verification script to confirm required design evidence exists before delivery.
| sac-y/identity-skill | chubbyguan/chubbyskills | ohad6k/emulo | |
|---|---|---|---|
| Stars | 202 | 205 | 199 |
| Language | Python | Python | Python |
| Setup difficulty | easy | moderate | easy |
| Complexity | 3/5 | 3/5 | 2/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Installed via a single npx command, works best with Codex and optional companion image-generation skills.
Identity Skill is an installable Agent Skill, mainly built for OpenAI's Codex, that builds a personal website matching reference images you approve before any code is written. The README, written in Chinese with an English translation available, explains that most AI website builders jump straight into writing code, so you cannot see what the design looks like until it is finished, and the visual taste of the AI can be unreliable. This skill flips that order: it generates a reference image for each section of the site first, waits for you to confirm it, and only then builds a pixel-matching implementation. The process runs in five steps, each with a checkpoint where a person must approve before moving on. Step one reads your resume, LinkedIn, or project links to decide the site's structure and offers up to three style directions. Step two produces a reference image for each section and locks it with a saved, non-overwritable hash. Step three breaks each reference image down into a list of assets, measuring the size and position of each piece. Step four generates only the approved assets, each one checked against its reference region. Step five hands everything to a website-building tool for a faithful, section-by-section implementation, checked with desktop, ultra-wide, and mobile screenshots. Installation is a single command, npx skills add Sac-Y/identity-skill, after which a user can simply describe what they want in Codex, such as building a portfolio site from their LinkedIn profile. A bundled Python verification script checks that reference image hashes, screenshots, and other required evidence exist before the site is considered delivered, though the README notes that passing verification confirms the evidence exists, not that the final visual quality is good. The project depends on optional image-generation and frontend-building skills if they are already installed, and falls back to built-in versions otherwise. It is released under the MIT license.
An AI agent skill that builds a personal website by first generating and locking approved reference images per section, then implementing a pixel-matching site from them.
Mainly Python. The stack also includes Python, Agent Skills, Codex.
MIT license, free to use, modify, and distribute with attribution.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.