Skill
Give your AI assistant direct access to real-time HK/US stock data and trading capabilities with Longbridge Skill
Free agent skill packs by Liz — landing replication, Doubao TTS, video scripts, GEO job hunt, WebGL backgrounds, DESIGN.md. Unzip and use.
Translation pending
SkillsPython · TypeScriptGitHub @lizliz404OnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Downloadable agent skills for AI coding assistants — landing page replication, WebGL Three.js backgrounds, Doubao TTS & podcasts, video scripts, GEO job hunting, DESIGN.md visual systems. Unzip and use. MIT.
Taken from the repository README
Loading the full README. The saved excerpt is shown below.
Clone-and-use skill packs for AI coding assistants — Claude Code, Cursor, Codex, Hermes, or any agent that reads a SKILL.md. Thirteen packs in skills/, each just a folder. No installers, no frameworks, no signup.
Also published on lizliz.xyz/skills — same packs, browsable online.
Agent skills are instruction packs for coding AIs. Instead of re-explaining your process every session — "capture the page, check density, verify offline" — you hand the agent a skill folder and it follows the workflow, runs the scripts, and hits the same gates you would.Read the full README ↗Clone-and-use skill packs for AI coding assistants — Claude Code, Cursor, Codex, Hermes, or any agent that reads a SKILL.md. Fourteen packs in skills/, each just a folder. No installers, no frameworks, no signup.
Also published on lizliz.xyz/skills — same packs, browsable online.
graph TD
README["README.md — you are here"] --> SKILLS["skills/ · fourteen packs<br/>each: SKILL.md + optional scripts/ + references/ (+ evals/)"]
SKILLS --> VOICE["Voice & content<br/>doubao-tts · video-script-conversion · talkcut"]
SKILLS --> VISUAL["Replication & visual<br/>landing-page-replication-v5 · webgl-threejs-background-animation<br/>interactive-projects-stream · design-md-visual-system"]
SKILLS --> TOOLS["Data & tooling<br/>geo-job-hunt · seo-master · clerk-auth · feedback-pipeline"]
SKILLS --> NET["Network / infra<br/>ai-native-mihomo · ai-native-proof-ledger · cloudflare-cost-guard"]
SKILLS -.-> SITE["lizliz.xyz/skills"]
SKILLS -.->|"soft route"| DT["design-templates asset pack"]
Agent skills are instruction packs for coding AIs. Instead of re-explaining your process every session — "capture the page, check density, verify offline" — you hand the agent a skill folder and it follows the workflow, runs the scripts, and hits the same gates you would.
These packs come from real pipelines on lizliz.xyz: they were built to get actual work done, then packaged so anyone can reuse them. AI-Native Proof Ledger is a researched reference skill; Cloudflare Cost Guard combines official documentation with reviewed operational lessons; runtime coverage and delivery still require acceptance. Neither is presented as a production-validated pipeline. A skill is just a folder — SKILL.md (the workflow map) plus optional scripts/ and references/ that the agent loads when it needs them.
SKILL.md can follow the workflow; scripts are stdlib-only Python where possible.Pick one pack, or grab everything:
# Everything at once
git clone https://github.com/lizliz404/agent-skills.git
# ...then copy the pack folders you want into your agent's skills directory
# Or cherry-pick one pack straight into place
git clone --depth 1 https://github.com/lizliz404/agent-skills.git
cp -r agent-skills/skills/doubao-tts ~/.claude/skills/
Where "your agent's skills directory" lives:
| Agent | Path |
|---|---|
| Claude Code | ~/.claude/skills/ |
| Cursor | ~/.cursor/skills/ |
| Hermes | ~/.hermes/skills/ |
| Codex / others | check the agent's docs — most read SKILL.md from a skills folder |
Some packs need environment variables (AMAP_MAPS_API_KEY, MCP_LIEPIN_API_KEY, DOUBAO_API_KEY) — each pack's README explains what it needs and where to get it.
Point your agent at the skill and let it work:
Use the landing-page-replication-v5 skill to copy https://example.com — capture, audit, and report the fidelity gates.
用 geo-job-hunt 技能:在天河区半径 5 公里内找前端岗位,反向确认公司,列 10 个可投的。
Each pack's SKILL.md starts with a workflow map, so the agent knows what to do without you spelling it out every time.
Every pack follows the same shape — progressive disclosure:
| Piece | Purpose |
|---|---|
SKILL.md | The workflow map and rules — loaded when the skill is invoked |
scripts/ | Run when a step needs computation (capture, transcribe, count, apply) |
references/ | Deep notes loaded on demand: API details, cases, checklists |
evals/ | Where provided: behavior scenarios or executable checks; declared cases alone are not passing test evidence |
The agent reads the map first and pulls in only the files the current task needs.
This repo exists because of a few beliefs:
Created by @lizliz404.
templates/sibling-routes.mdWEBGL_THEATER or agent chrome is the mechanism — it does not vendor those folders.MIT — Use it, modify it, share it.