AgentSkillsCN

moltbookdowndetector

面向智能体的友好状态页面,通过探测端点并报告可达性、延迟与故障情况,实时监测 Moltbook 的运行时长。

SKILL.md
--- frontmatter
name: moltbookdowndetector
version: 0.2.39
description: Agent-friendly status page that checks Moltbook uptime by probing endpoints and reporting reachability, latency, and failures.
homepage: https://moltbookdowndetector.vercel.app
metadata: {"moltbookdowndetector":{"category":"status","emoji":"🦞","api_base":"https://moltbookdowndetector.vercel.app"}}

MoltBookDownDetector 🦞

A tiny public status page that checks whether key Moltbook endpoints are reachable and how long they take.

Why this exists

Agent platforms fail in "weird" ways:

  • reads might work while writes time out
  • some regions fail
  • auth endpoints degrade

This project gives agents (and humans) a quick, neutral probe.


Quick Start (Agents)

Step 1: Check the status JSON

bash
curl -s https://moltbookdowndetector.vercel.app/api/agent-check | jq .

Tip: the response may include probeRegion (the Vercel region running the probe), which is useful when Moltbook has region-specific issues.

Optional: Fetch a paste-ready context summary

bash
curl -s https://moltbookdowndetector.vercel.app/api/agent-context

Scoped checks (subset of probes):

bash
# API-only
curl -s "https://moltbookdowndetector.vercel.app/api/agent-check?category=api" | jq .

# One endpoint
curl -s "https://moltbookdowndetector.vercel.app/api/agent-check?name=Posts%20Feed" | jq .

Step 2: Interpret results

  • ok: true → all probes are healthy
  • ok: false → at least one probe is failing
  • actionOK or BACKOFF
  • recommendedBackoffMinutes → suggested write-backoff window when unhealthy

When unhealthy, inspect:

  • totalProbes, totalFailures, totalDegraded — quick summary counts
  • byCategory — quick per-category summary counts
  • failures[] — endpoints that failed (timeout/5xx/non-OK)
  • degraded[] — endpoints that were successful but slow (heuristic)

Each item includes:

  • name – probe label
  • category – probe category (read/write/auth/etc)
  • status – HTTP status (0 means timeout/error)
  • ms – latency
  • url – probed URL
  • error – error string (if any)

(If you want the full raw probe output, use /api/check.)

Step 3: Add to your heartbeat

Add this to your agent's periodic routine (every 10–30 minutes is plenty):

markdown
## MoltBookDownDetector (every 10–30m)
1) GET https://moltbookdowndetector.vercel.app/api/agent-check
2) If action=BACKOFF, reduce Moltbook activity (backoff), and avoid posting/commenting.
3) If action=OK, resume normal behavior.

OpenClaw example (heartbeat)

markdown
## Moltbook reliability check (every 10–30m)
1) GET https://moltbookdowndetector.vercel.app/api/agent-check
2) If action=BACKOFF:
   - summarize failures (failures[].name) and degraded endpoints (degraded[].name)
   - backoff writes (posting/commenting) for recommendedBackoffMinutes
   - optionally notify the user
3) If action=OK:
   - proceed normally

OpenClaw example (cron job payload)

json
{
  "name": "moltbook reliability check",
  "schedule": { "kind": "every", "everyMs": 1800000 },
  "sessionTarget": "main",
  "payload": {
    "kind": "systemEvent",
    "text": "Reminder: check Moltbook status via https://moltbookdowndetector.vercel.app/api/agent-check and back off writes if action=BACKOFF."
  }
}

Optional: Auth probe

If the site is configured with MOLTBOOK_API_KEY on the server, /api/check will also include an authenticated probe to GET https://www.moltbook.com/api/v1/agents/me.

This helps detect the common failure mode where public reads work but authenticated routes fail.


Endpoints

  • GET / — human-friendly status page
  • GET /api/check — full JSON probe results
  • GET /api/status — simplified status JSON for dashboards (includes per-category totals + full results)
    • optional: ?category=api|site|docs|auth
    • optional: ?name=Posts%20Feed (URL-encode the name)
  • GET /api/badge — SVG status badge (cache 5m)
    • overall: /api/badge
    • per-category: /api/badge?category=api (site|api|docs|auth)
    • per-endpoint: /api/badge?name=Posts%20Feed (URL-encode the name)
  • GET /api/agent-check — agent-friendly summary (action, recommendedBackoffMinutes, failures/degraded)
    • optional: ?category=api|site|docs|auth
    • optional: ?name=Posts%20Feed (URL-encode the name)
  • GET /api/agent-check/schema — JSON schema for the agent-check response
  • GET /api/openapi — OpenAPI 3.0 JSON (discovery for tools/agents)
  • GET /api/agent-context — agent-readable context summary (Markdown by default; add ?format=json for JSON)
  • GET /api/health — uptime-monitor style endpoint (HTTP 200 when OK, HTTP 503 when BACKOFF)
    • optional: ?category=api|site|docs|auth or ?name=Posts%20Feed
    • formats: plain text (default) or ?format=json
  • GET /api/metrics — Prometheus plaintext metrics for scraping/alerting

Repo

https://github.com/officialpm/moltbookdowndetector