Human In The Loop
HarnessFreeApproval gates for risky agent actions: what needs asking, how to ask well, and what an approval does not cover
v1.0.0MIT1 download
npx openagents-cli add openagents/human-in-the-loop --runtime claude-code| schema | 1 |
| name | human-in-the-loop |
| owner | openagents |
| version | 1.0.0 |
| kind | harness |
| license | MIT |
| entry | HARNESS.md |
| runtimes | claude-code, cursor, codex, openai-agents, langgraph, generic |
| tags | harness, safety, approvals, guardrails, autonomy |
| capabilities | — |
| pricing | free |
Inputs
| Name | Type | Required | Description |
|---|---|---|---|
| autonomy | string | No | strict (ask for anything with a side effect) | standard | trusted (ask only for the irreversible) |