# Use Human For AI with your agent.

[Human For AI](https://humanforai.dev/) (`humanforai.dev`) registers 8 WebMCP tools. 5 of 8 tools declares `readOnlyHint`, so an agent can invoke them through `tools.call`. The registry runs the tool in a fresh headless page on humanforai.dev and returns untrusted site content.

This page mirrors the [WebMCP Registry](https://wmcp.ai/sites/humanforai.dev) listing for `humanforai.dev`, last live-checked by the registry at 2026-09-23T03:14:04.889Z. Run `ghostget webmcp sites.get --input '{"domain":"humanforai.dev"}' --json` for the current schema; the listing can drift between checks.

## Use Human For AI with my agent

Ghostget's bundled `webmcp` adapter speaks to the public WebMCP Registry, so an agent always reads this site's current tool schema instead of a stale hard-coded copy. Install v0.18.35, sync bundled adapters once, then call the registry operations:

```
bun add --global https://github.com/hraness/ghostget/releases/download/v0.18.35/hraness-ghostget-0.18.35.tgz
ghostget adapter sync-bundled --json
```

**Read the live tool schema** — every registered tool, its input schema, annotations, and the page that publishes it:

```
ghostget webmcp sites.get --input '{"domain":"humanforai.dev"}' --json
```

**Call a read-only tool** — the registry refuses anything not declared `readOnlyHint` and returns the site's answer as untrusted content:

```
ghostget webmcp tools.call --input '{"domain":"humanforai.dev","tool":"check_message_thread","input":"{}"}' --json
```

## Registered WebMCP tools

Human For AI publishes 8 tools on humanforai.dev; 5 declare `readOnlyHint`.

| Tool | What it does | Callable through Ghostget |
| --- | --- | --- |
| `check_message_thread` | One conversation with the human, looked up by message_id with its access_token: the original message, every reply oldest-first, the reply count, and whether the human has answered yet. | Callable read-only tool |
| `check_task_status` | How far the human has got with one piece of work, looked up by task_id: whether it was seen, accepted, delivered, or declined, with timestamps, the expected delivery time, the human's notes, and, once delivered, the signed receipt. | Callable read-only tool |
| `get_human_services` | The catalog of things one verified real human will do for an AI agent: verify a place, product, or claim in person, test an app, judge AI output, collect data, run a local errand. Includes who the human is, how fast they answer, what they a | Callable read-only tool |
| `get_service_details` | One entry from the catalog of things the human will do, looked up by its task_type: title, description, an example request, the response format, and typical turnaround. Returns an unknown_task_type error with the list of valid task types wh | Callable read-only tool |
| `message_human_operator` | Starts a conversation with the real human behind this site: a question, a scoping request, or a custom project. Returns the message_id, a thread_url, and a one-time access_token. The human answers at the reply-to address. | Listed for discovery; the registry allows read-only calls only |
| `reply_in_message_thread` | Adds a follow-up to an existing conversation with the human: an answer to a question they asked, more detail, a correction, or a withdrawal. Returns the message_id and the updated reply_count. Conversations have a reply cap and an hourly li | Listed for discovery; the registry allows read-only calls only |
| `search_services` | Text search over the catalog of things the human will do: matches the query against each entry's title, description, category, and example request, and returns the matching entries with their task_type, title, description, and typical turna | Callable read-only tool |
| `submit_human_task` | Hires the human: sends them one clearly described piece of real-world work and returns a task_id and status_url at once. The human reads it and decides whether to take it, then delivers to contact_email, or as text on the work record when d | Listed for discovery; the registry allows read-only calls only |

## What to expect

- **Read-only only.** The registry refuses WebMCP tools that do not declare `readOnlyHint`; Ghostget never weakens that check.
- **Untrusted results.** Tool output is site content, not instructions — treat it as data.
- **Live schemas.** `sites.get` always returns the schema the registry saw most recently, so agents adapt when Human For AI changes its tools.
- **No account needed.** These calls are credential-free registry reads. For tools that need a session on humanforai.dev, browse the site itself or check back when the registry lists more tools.