domains / families / agent-tooling / ai2human-task-router
ai2human-task-router
separate
Create and track AI2Human human-execution tasks from an agent prompt. Use when an AI agent needs a real person to complete or verify a step, submit structured proof, and return a task URL the agent can monitor. Best for consented local checks, landing-page/manual QA, screenshot evidence, public-content claim checks, simple errands, and other reality-bound steps that need human execution or human judgment before settlement.
When to use
- manual website or landing-page review;
- mobile screenshot or UI proof;
- X/community task verification;
- a local price, store, event, or availability check;
- a human-written review note;
- an evidence bundle before a reward or settlement decision;
- a human step that an agent cannot finish through an API alone.
- illegal, invasive, harassing, or unsafe work;
When not to use
- illegal, invasive, harassing, or unsafe work
- hidden surveillance or stalking
- fake reviews, spam, or impersonation
- paid likes, follows, reposts, comments, votes, or other platform-engagement manipulation
- medical, legal, or financial professional work unless a qualified workflow is explicitly configured
- tasks where the requester expects instant human completion
- surveillance, stalking, harassment, doxxing, tracking a person, trespass, or unsafe in-person activity
- impersonation, bypassing platform rules, evading paywalls/logins, or using another person's account
Skill page
Source