AgentSkillsCN

paid-agent

针对电力系统开展直流潮流分析。当需要利用直流近似法计算功率潮流、构建导纳矩阵、计算线路潮流与负载百分比,或对输电网络进行灵敏度分析时,这一技能将大显身手。

SKILL.md
--- frontmatter
name: paid-agent
description: |
  Builds and deploys a production-ready paid Lucid Agent from user requirements.
  Scaffolds with CLI, implements entrypoints, reviews code, tests, publishes to GitHub, and deploys to Railway.
  Use when: "create a paid agent for X", "build and deploy an agent", "make a monetized agent".
  Not for: automated topic discovery (agent-factory) or platform-hosted JS handlers (lucid-agent-creator).
allowed-tools: [Skill, Bash, Read, Write, Edit, AskUserQuestion]

Paid Agent Pipeline

Step 1: Gather Requirements

Ask the user for:

  • What the agent does
  • Agent name (kebab-case)
  • Endpoints — which are free vs paid, and pricing
  • Deploy now?

Step 2: Scaffold

bash
bunx @lucid-agents/cli {agent-name} --adapter=hono --non-interactive
cd {agent-name}
bun install

Step 3: Implement Entrypoints

Edit src/index.ts to add entrypoints. Use lucid-agents-sdk skill for API patterns.

At least 1 free + N paid endpoints. All must return real data.

Step 4: Test

bash
bun test
bun run dev
curl -X POST http://localhost:3000/entrypoints/{key}/invoke \
  -H 'Content-Type: application/json' -d '{"input": { ... }}'

Step 5: Code Review

bash
Skill("feature-dev:code-reviewer")

Fix issues, re-run tests.

Step 6: Push to GitHub

bash
git init
git add src/ package.json tsconfig.json .gitignore README.md
Skill("commit")
gh repo create {agent-name} --public --source=. --remote=origin --push --description "{description}"

Step 7: Deploy (Optional)

bash
Skill("railway-deploy")

Step 8: Summary

Report to user: GitHub URL, Railway URL (if deployed), endpoint table with prices, curl examples.