A2A (Agent2Agent)

Let other AI agents delegate monitoring tasks to WatchFor over the Agent2Agent protocol — diagnose incidents, check status, and take operator actions.

WatchFor is an A2A (Agent2Agent) agent, so other AI agents can delegate monitoring tasks to it — check status, diagnose what's broken and why, report on incidents and reliability, and take operator actions — without a bespoke integration.

A2A is the horizontal counterpart to our MCP server: MCP connects an agent to WatchFor as a tool; A2A lets an autonomous agent hand WatchFor a task and get a result back.

Endpoint: https://watchfor.io/api/a2a (JSON-RPC 2.0) Agent Card: /.well-known/agent-card.json

Authentication

Same credentials as the REST API and MCP server (details): a wf_live_… API key or an OAuth 2.1 access token, passed as a Bearer header. Read skills need the read scope; the write skills (create monitor, acknowledge/resolve incident, schedule maintenance) need write. The Agent Card advertises both schemes so OAuth-capable clients can discover and connect on their own.

Skills

WatchFor is a deterministic skill executor: a peer agent selects a skill by id and passes its input; WatchFor runs it and returns a completed task with a short human summary plus the structured facts behind it.

The 13 skills are deliberately task-shaped, not raw CRUD: they cover understanding your monitoring and taking the operator actions an agent realistically performs mid-task. Full administrative CRUD (editing alert-rule thresholds, managing contacts, billing) lives on the REST API and MCP server.

Understand & report (read):

Skill	What it answers / does
`list-monitors`	What is being watched — id, name, type, status
`check-status`	Is it up right now? (one monitor or the whole org)
`diagnose-incidents`	What is broken and why — every active incident, what failed, category, duration
`explain-monitor`	Describe one monitor: what it watches, health, uptime, alert rules, recent checks & incidents
`incident-report`	What happened in one incident — trigger + full timeline
`list-incidents`	Incident history with filters (status, severity, monitor, time range)
`reliability-report`	Incident counts and MTTR over a period
`describe-monitor-types`	What can be monitored + each type's config fields and alert metrics
`list-locations`	Available probe locations (ids for create-monitor)

Act (write):

Skill	What it does
`create-monitor`	Start watching a target
`manage-monitor`	Pause / resume / run-check / update / delete a monitor
`manage-incident`	Acknowledge or resolve an incident
`manage-maintenance`	Create / list / cancel maintenance windows (e.g. silence alerts during a deploy)

Calling a skill

Send a JSON-RPC message/send whose message carries a data part selecting the skill and its input:

curl https://watchfor.io/api/a2a \
  --header "Authorization: Bearer wf_live_YOUR_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "message/send",
    "params": {
      "message": {
        "role": "user",
        "messageId": "1",
        "parts": [
          { "kind": "data", "data": { "skill": "diagnose-incidents", "input": {} } }
        ]
      }
    }
  }'

The response is a completed A2A task; the first artifact part is a plain-language summary and the second is the structured data:

{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "kind": "task",
    "status": { "state": "completed" },
    "artifacts": [
      {
        "name": "diagnose-incidents-result",
        "parts": [
          { "kind": "text", "text": "2 active incidents: api.example.com — HTTP status code (…)" },
          { "kind": "data", "data": { "active_count": 2, "incidents": [ /* … */ ] } }
        ]
      }
    ]
  }
}

If you omit the skill selector, the task comes back input-required with the list of available skills — so an agent can discover them at runtime.

Notes

Synchronous. Every skill completes inside message/send; there are no long-running tasks to poll and message/stream is not offered (capabilities.streaming = false).
Behaviour is the API. Each skill calls /api/v1 with your credential, so scopes, per-plan rate limits and validation apply identically.

A2A (Agent2Agent)

Authentication

Skills

Calling a skill

Notes

On this page