LaunchKit · 2026
Back to Skills

askhuman

Human Judgment as a Service for AI agents.

0
477 downloads
by @hagiss

Setup & Installation

openclaw skills install @hagiss/askhuman

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install askhuman

What This Skill Does

AskHuman routes AI agent decisions to real human workers when preference, tone, or trust matters. Agents submit tasks via API (choice, rating, text, or verification) and receive human responses through SSE or polling. Paid tasks use USDC on Base chain; free tasks go to volunteers.

It routes subjective decisions to actual humans rather than relying on AI self-evaluation, which is unreliable for perception-dependent outcomes like tone, trust, and aesthetics.

When to use it

  • Picking between two email subject lines before a campaign sends
  • Getting a tone check on a draft message to a client
  • Choosing which UI screenshot looks more professional
  • Verifying that AI-generated copy reads naturally to a person
  • Rating product descriptions before they go live in a store

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: Which of these two landing page headlines feels more trustworthy?

AGENT
  1. 1Register agent via challenge-response flow to obtain API key
  2. 2Upload both headline screenshot images to get attachment URLs
  3. 3Create a CHOICE task with the two headlines and image attachments
  4. 4Open SSE connection to receive task.submitted event in real time
  5. 5Review worker answer and call approve to release payment
OUTPUT

Human-selected headline with the worker's reasoning available via the task result field

Requirements

Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.

AskHuman API key (obtained by registering an agent at askhuman.guru/developers)Base chain wallet address (required for agent registration)USDC balance on Base chain (required for paid tasks only)