ams-trigger-dev
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-11-07
- Stars
- ★ 0
- Forks
- 1
03 SKILL.md 正文
Trigger.dev
Overview
Trigger.dev is an open-source background job platform for TypeScript: define tasks in code, trigger via events/schedules, get retries and observability built in.
It solves the classic pain: background work (emails, webhooks, data syncs, AI pipelines) that's reliable, visible, and retryable — without building job infrastructure.
Model: tasks as code, runs as durable executions, dashboard for monitoring, schedules and event triggers built in.
When to use
- Adding background jobs to a TypeScript/Next.js app
- Replacing fragile cron scripts or queue DIY
- Scheduled tasks (reports, syncs, cleanups) with monitoring
- Event-driven workflows (on signup -> onboard sequence)
- Long-running tasks (AI pipelines, video processing, bulk ops)
Core concepts
- Tasks as code. Define tasks with schemas (e.g., zod) for type-safe payloads. Tasks live with your app code — versioned, reviewed, tested like the rest.
- Triggers. Event triggers (from your app), schedules (cron), and manual. One task, multiple trigger sources — reuse the logic.
- Retries and idempotency. Automatic retries with backoff; design task bodies idempotent (safe on retry). Use idempotency keys for external side effects.
- Concurrency and queues. Control parallelism per task; queues for ordering and rate-limiting. Protect downstream APIs from thundering herds.
- Runs dashboard. Every execution visible: inputs, logs, timing, retries. Debugging production jobs without SSH-ing into servers.
- Schedules. Cron-like scheduling with timezone support. Replace scattered cron jobs with visible, monitored scheduled tasks.
- Environments. Dev/staging/prod separation for tasks. Test tasks safely before they touch production data.
- Self-hosting option. Open source — self-host for data control. Managed cloud for zero ops. Choose by your constraints.
Practical workflow
- Identify background work. List: what runs async today (crons, scripts, queue hacks)? What should? Prioritize by pain (failures, invisibility).
- Define tasks.
One task per job with typed payload schema. Name clearly (
send-welcome-email,sync-crm-contacts). Keep tasks focused. - Wire triggers. Events from app code, schedules for periodic work. Start with one trigger per task; add more as patterns emerge.
- Build idempotent bodies. Each task: validate input, check-before-act where possible, external calls with idempotency keys, structured logging.
- Configure retries/queues. Retry policy per task criticality; concurrency limits protecting downstream APIs; queues where ordering matters.
- Test in dev. Run tasks locally/with dev environment using real-shaped payloads. Verify retries by simulating failures.
- Deploy and monitor. Ship with dashboard bookmarks; set up failure alerts to a watched channel. First week: watch runs closely.
- Migrate crons gradually. Move one cron/script at a time. Keep old and new running in parallel briefly; verify parity, then cut over.
Common pitfalls
- Non-idempotent tasks. Retries creating duplicate emails/charges/records. Idempotency isn't optional with automatic retries.
- No payload validation. Untrusted event data crashing tasks. Schema-validate at the boundary, every time.
- Unbounded concurrency. 1,000 parallel tasks hammering a rate-limited API. Concurrency limits and queues exist — use them.
- Ignoring the dashboard. Failures visible in the dashboard, nobody watching. Alerts to a real channel, reviewed daily.
- Cron sprawl migration. Moving 30 crons at once. One at a time, verified, or you'll debug 30 new failure modes simultaneously.
- Secrets in task code. API keys hardcoded in task definitions. Environment secrets, never in code or logs.
- No dev/prod separation. Testing tasks against production data/APIs. Environments exist for a reason — use them.
- Fire-and-forget events. Triggering without confirming the task ran. For critical flows, verify completion or alert on absence.
04 同仓库更多 Skills
web-scrape
Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open/fetch fails on JS-heavy pages, or when you need clean markdown/structured extraction from a site. Requires FIRECRAWL_API_KEY.
design-taste-frontend
tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML/CSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, "AI-made", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.
test-driven-development
Use when implementing any feature or bugfix, before writing implementation code
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
requesting-code-review
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).