360 Secure SkillHub
360 安全技能中心

HookPulse

该 Skill 借助 HookPulse 服务监控定时任务、webhook、备份和计划执行的 GitHub Actions 工作流：创建按间隔或 cron 调度的监控、导入 crontab、上报开始/成功/失败信号与退出码，读取监控状态与公开状态页；漏跑或失败时通过邮件或 webhook 告警，恢复后发送恢复消息。

wrpbr
wrpbr
热度 36.68
最近更新 2026-10-15 11:13:17
分类：自动化依赖：MCP 服务 / 第三方 API / 代码执行应用场景：工作流 / 自动化脚本 / DevOps / 日程安排

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add wrpbr/hookpulse
Metadata
name
hookpulse
description
Monitor cron jobs, webhooks, backups and scheduled GitHub Actions workflows with HookPulse (hookpulse.net), a dead man's switch. Create a monitor with an interval or a cron schedule, import a crontab, send start, success and failure signals with the exit code, instrument a fleet with one setup key, and read monitor state, check-ins and the public status page. A missed or failed run alerts by e-mail or webhook. Use when a task mentions a cron job, crontab, scheduled task, heartbeat, dead man's switch, missed run, backup that stopped, or a GitHub Actions schedule.
license
MIT-0
compatibility
Needs HTTPS access to hookpulse.net. Works with the remote MCP server or with plain HTTP calls.

HookPulse

HookPulse watches routines that should run on a schedule. The routine sends a check-in to its link when it runs. If the check-in is late, or the routine reports a failure, HookPulse sends an alert. The next good check-in sends a recovery message.

Connect

  • MCP server (preferred): https://hookpulse.net/mcp?checkout=site. Transport: Streamable HTTP. Every MCP tool is free, within its free limits: no payment.
  • HTTP API: base URL https://hookpulse.net. Each MCP tool calls one route of this API.
  • Identity: create_guest returns a guest token hp_…. Send it as X-Guest-Token on later calls. Keep it: the monitors belong to it until the person signs in on the site.
  • Discovery: GET https://hookpulse.net/api/ lists the routes, prices and limits. https://hookpulse.net/llms-full.txt has the full reference.

MCP tools

Tool HTTP Use it to
create_guest POST /api/guest Get the guest token that owns the monitors.
create_endpoint POST /api/endpoints Create a monitor: interval_sec, or cron + tz + grace_sec, and the alert destinations.
import_crontab POST /api/import/crontab Read a crontab and propose one monitor per line. Nothing is written without apply=1.
list_endpoints GET /api/endpoints List the monitors and their state.
get_endpoint GET /api/endpoints/{id} Read one monitor: state, next expected time, last check-in.
list_events GET /api/endpoints/{id}/events Read the latest check-ins of a monitor.
delete_endpoint DELETE /api/endpoints/{id} Remove a monitor.
ping_ingest GET /in/{id} Send a success check-in, as the routine does.
signal_start GET /in/{id}/start Record that a run started, to measure its duration.
signal_fail GET /in/{id}/fail Record a failed run; the alert leaves on the next five-minute check.
signal_exit_code GET /in/{id}/{exit_code} Record the shell exit code: 0 is success, 1 to 255 a failure.
ingest_key GET /api/ingest-key Get the setup key for /in/{key}/{slug}: the first check-in of a name creates its monitor.
ping_slug GET /in/{key}/{slug} Check in by name with the setup key, creating the monitor the first time.
status_feed_url GET /api/status-feed Get the public status page and feed of all monitors of the owner.
list_templates GET /api/templates Get ready snippets: curl, crontab line, n8n, GitHub Actions and the alert JSON.

Rules

  1. If an MCP tool answers free_limit, tell the user that the free limit for that action is used up. Before you create more than ten monitors, read the free allowances at GET /api/billing on the HTTP API. Do not use an allowance from memory.
  2. Never pay by yourself. On the HTTP API, a paid route returns HTTP 402: show the price and https://hookpulse.net/pricing to the user. The person buys on the site.
  3. The monitor ID is a credential: whoever has it can send check-ins and silence the alarm. Do not print it in public logs or issues. The status page link is the one to share.
  4. A start or failure signal is not proof of life. Only a success check-in resets the expected time.
  5. For a schedule, prefer cron with tz and a grace_sec that covers the usual delay and run time.
  6. If you get HTTP 429 or 503, wait for the time in Retry-After and try once more.

Examples

# Guest token, then a monitor for a backup at 03:00 São Paulo time with 15 minutes of allowed delay
HP=$(curl -s -XPOST https://hookpulse.net/api/guest | jq -r .token)
curl -s -XPOST https://hookpulse.net/api/endpoints -H "X-Guest-Token: $HP" -H 'content-type: application/json' \
  -d '{"name":"nightly-backup","cron":"0 3 * * *","tz":"America/Sao_Paulo","grace_sec":900}'

# At the end of the routine: the exit code of the last command
curl -fsS "https://hookpulse.net/in/$ID/$?"

Limits