Skip to main content

The /hai-agents skill writes and runs a TypeScript availability check from one prompt

The hai-agents skill teaches your coding assistant H’s APIs, so it scaffolds working sessions, agents, and environments from a plain-language prompt instead of guessing at endpoints. One SKILL.md works in Claude Code, Cursor, and Hermes. Only the install differs.

hcompai/computer-use-agents-demos/skills/hai-agents

The skill’s source: SKILL.md plus the reference docs it loads. Also published as a Claude Code plugin (marketplace hai-skills).

​
What your assistant learns

  • Portal: auth, organizations, invitations, billing, API keys, and a login script that writes your key into .env.
  • Agents API: sessions, agents, skills, environments, vaults, and /changes long-polling.
  • SDKs: the Python and TypeScript clients (run_session / runSession, session handles, events, error classes).
  • Agent View: replaying and sharing what an agent did.
The skill triggers on its own when the conversation matches: asking about H sessions or hk-... keys pulls it in.

​
Install

Add the marketplace first (otherwise hai-skills is unknown), then install. Type these in any Claude Code prompt:

​
Build a use case

Describe the workflow in plain language. In Claude Code you can call the plugin’s slash command. In Cursor or Hermes, a plain prompt triggers the skill.
❯ /hai-agents:hai-agents “Add one iPhone 17 Pro to my Amazon.com cart”
It writes TypeScript or Python against the real endpoints and SDK helpers. The video above came from:
❯ /hai-agents:hai-agents “Generate TypeScript code that navigates to jacquemus.com, finds the France × Nike football jersey, and checks its availability in sizes S and XXL.”
Generated code: examples/product_availability/src/index.ts. More recipes (QA via CLI, schema-driven extraction, counterfeit detection) are in the demos repo. Related: Quickstart, SDKs, MCP server