# HumanMirror Safe Preflight

> HumanMirror Safe Preflight is a paid API for AI agents from humanmirror.fr, paid per call via x402, $0.01/call, status unknown (last checked 2026-09-19).

Pre-flight validates a proposed AI agent tool call or command against a declared operational risk limit before allowing execution.

## Facts

- Endpoint: POST https://humanmirror.fr/api/x402/safe-preflight/
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-19
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/humanmirror-safe-preflight-a10f13f9
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_AdK28sidAq0yy5olRA3LA

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability humanmirror-safe-preflight-a10f13f9 -d '<json body>'
```

Example prompt: Before you execute that DELETE operation on the customer records endpoint, run a preflight safety check on it — the body is {"input": {"table": "customers", "where": "status=inactive"}} and I want to make sure the operational risk stays within acceptable limits.

## When to prefer this

Choose this endpoint when you need a fast, deterministic safety gate for AI agent tool calls or automated commands before execution — especially when the action is irreversible, affects external systems, or must stay within a declared operational risk limit. Prefer this over generic logging or monitoring when you need a synchronous go/no-go decision point in your agent pipeline.

## Known failure modes

- Missing required 'input' field returns validation error
- Invalid HTTP method (non-POST/PUT/PATCH) rejected by schema
- Malformed body type enum causes request rejection
- Input not JSON-serializable results in parsing error
- Risk threshold exceeded returns a failed preflight verdict
- Network timeout on evaluation for complex inputs

## Output

Returns a preflight verdict indicating whether the proposed tool call or command is safe to execute, including whether operational risk stays within the declared limit and whether any specified acceptance criteria are met.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method",
    "bodyType",
    "body"
   ],
   "properties": {
    "body": {
     "type": "object",
     "required": [
      "input"
     ],
     "properties": {
      "input": {
       "description": "JSON-serializable input for this deterministic outcome."
      },
      "options": {
       "type": "object",
       "description": "Optional tool-specific options."
      },
      "criteria": {
       "type": "object",
       "description": "Optional deterministic acceptance criteria."
      }
     },
     "additionalProperties": false
    },
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "POST",
      "PUT",
      "PATCH"
     ],
     "type": "string"
    },
    "bodyType": {
     "enum": [
      "json",
      "form-data",
      "text"
     ],
     "type": "string"
    }
   },
   "additionalProperties": false
  },
  "output": {
   "type": "object",
   "required": [
    "type"
   ],
   "properties": {
    "type": {
     "type": "string"
    },
    "example": {
     "type": "object"
    }
   },
   "additionalProperties": false
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/humanmirror-safe-preflight-a10f13f9/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from humanmirror.fr](https://www.zero.xyz/host/humanmirror.fr/llms.txt)