pp-trigger-dev

What this skill does

<!-- GENERATED FILE — DO NOT EDIT.
     This file is a verbatim mirror of library/developer-tools/trigger-dev/SKILL.md,
     regenerated post-merge by tools/generate-skills/. Hand-edits here are
     silently overwritten on the next regen. Edit the library/ source instead.
     See the repository agent guide, section "Generated artifacts: registry.json, cli-skills/". -->

# Trigger.dev — Printing Press CLI

## Prerequisites: Install the CLI

This skill drives the `trigger-dev-pp-cli` binary. **You must verify the CLI is installed before invoking any command from this skill.** If it is missing, install it first:

1. Install via the Printing Press installer. It defaults binaries to `$HOME/.local/bin` on macOS/Linux and `%LOCALAPPDATA%\Programs\PrintingPress\bin` on Windows:
   ```bash
   npx -y @mvanhorn/printing-press-library install trigger-dev --cli-only
   ```
2. Verify: `trigger-dev-pp-cli --version`
3. Ensure the reported install directory is on `$PATH` for the agent/runtime that will invoke this skill.

If the `npx` install fails (no Node, offline, etc.), fall back to a direct Go install (requires Go 1.26.3 or newer):

```bash
go install github.com/mvanhorn/printing-press-library/library/developer-tools/trigger-dev/cmd/trigger-dev-pp-cli@latest
```

If `--version` reports "command not found" after install, the runtime cannot see the binary directory on `$PATH`. Do not proceed with skill commands until verification succeeds.

## When to Use This CLI

Reach for this CLI when an agent task needs to operate Trigger.dev programmatically: triaging failed runs, auditing schedule health, diffing env vars across environments, computing LLM costs across runs, or watching for new failures. It complements the official trigger.dev CLI (which focuses on dev-time deploy/init/dev) by exposing the management API surface as scriptable, agent-native commands with offline FTS and typed exit codes.

## Unique Capabilities

These capabilities aren't available in any other tool for this API.

### Real-time terminal alerting
- **`runs watch`** — Watches runs for new failures and interrupts your terminal the moment one happens — desktop notification, sound, and a non-zero exit so it composes with shell loops.

  _Reach for this when an agent or oncall engineer needs to react to failures the second they appear, instead of polling the dashboard or waiting for an email._

  ```bash
  trigger-dev-pp-cli runs watch --task daily-digest --notify --sound
  ```

### Local state that compounds
- **`schedules stale`** — Lists schedules that stopped firing or whose recent runs have a low success rate — the zombie cron audit no other tool gives you.

  _Reach for this after an env var rotation, a Postgres URL change, or any silent-config event that might have killed crons without anyone noticing._

  ```bash
  trigger-dev-pp-cli schedules stale --days 7 --min-success-rate 0.5 --agent
  ```
- **`runs span-cost`** — Walks recent runs, surfaces the most expensive LLM spans grouped by model and task, with token totals and dollar cost — the report your finance lead actually wants.

  _Reach for this when an AI-product engineer or agent needs to pinpoint which model + task pair is eating the LLM budget, before the next billing cycle closes._

  ```bash
  trigger-dev-pp-cli runs span-cost --since 7d --by model,task --top 20 --agent --select spans.model,spans.task_identifier,spans.total_cost_cents
  ```
- **`failures top`** — Top recurring (task, error-signature) pairs over a time window — mechanical group-by, no LLM, no NLP.

  _Reach for this in an incident loop when an agent needs to find the dominant failure signature instead of reading 200 stack traces by hand._

  ```bash
  trigger-dev-pp-cli failures top --since 7d --top 20 --agent --select patterns.task_identifier,patterns.error_signature,patterns.count,patterns.last_occurred_at
  ```
- **`runs find`** — FTS5 grep over the synced runs table — error messages, tags, metadata, task identifiers — ranked by recency.

  _Reach for this when an agent has a fragment of an error message and needs the matching runs without typing TRQL._

  ```bash
  trigger-dev-pp-cli runs find "payload too large" --status FAILED --since 30d --agent
  ```

### Operator hygiene
- **`envvars diff`** — Side-by-side diff of environment variables between two environments — keys missing, keys extra, values that differ (masked).

  _Reach for this when an agent is debugging "works in staging, fails in prod" — the answer is almost always an env var, and diff makes the answer one command away._

  ```bash
  trigger-dev-pp-cli envvars diff --from prod --to staging --project proj_abc --agent
  ```

### Agent ergonomics
- **`runs get`** — runs get returns typed exit codes: 0=COMPLETED, 20=FAILED, 21=CRASHED, 22=SYSTEM_FAILURE, 23=CANCELED, 3=not-found, 4=auth-error. Cobra annotation pp:typed-exit-codes makes verify and agents read the contract directly.

  _Reach for this when an agent is writing a shell loop over many runs and wants to branch on success/failure without parsing JSON._

  ```bash
  trigger-dev-pp-cli runs get run_abc123 --json && echo COMPLETED || echo $?
  ```

## Command Reference

**batches** — Manage batches

- `trigger-dev-pp-cli batches <batchId>` — Retrieve a batch by its ID, including its status and the IDs of all runs in the batch.

**deployments** — Manage deployments

- `trigger-dev-pp-cli deployments get-latest-v1` — Retrieve information about the latest unmanaged deployment for the authenticated project.
- `trigger-dev-pp-cli deployments get-v1` — Retrieve information about a specific deployment by its ID.
- `trigger-dev-pp-cli deployments list-v1` — List deployments for the authenticated environment, ordered by most recent first.

**projects** — Manage projects


**query** — Manage query

- `trigger-dev-pp-cli query execute-v1` — Execute a TRQL (Trigger.dev Query Language) query against your run data. TRQL is a SQL-style query language that...
- `trigger-dev-pp-cli query get-schema-v1` — Get the schema for TRQL queries, including all available tables, their columns, data types, descriptions, and...
- `trigger-dev-pp-cli query list-dashboards-v1` — List available built-in dashboards with their widgets. Each dashboard contains pre-built TRQL queries for common...

**queues** — Manage queues

- `trigger-dev-pp-cli queues list-v1` — List all queues in your environment with pagination support.
- `trigger-dev-pp-cli queues retrieve-v1` — Get a queue by its ID, or by type and name.

**runs** — Manage runs

- `trigger-dev-pp-cli runs list-v1` — List runs in a specific environment. You can filter the runs by status, created at, task identifier, version, and more.
- `trigger-dev-pp-cli runs retrieve-v1` — Retrieve information about a run, including its status, payload, output, and attempts. If you authenticate with a...

**schedules** — Manage schedules

- `trigger-dev-pp-cli schedules create-v1` — Create a new `IMPERATIVE` schedule based on the specified options.
- `trigger-dev-pp-cli schedules delete-v1` — Delete a schedule by its ID. This will only work on `IMPERATIVE` schedules that were created in the dashboard or...
- `trigger-dev-pp-cli schedules get-v1` — Get a schedule by its ID.
- `trigger-dev-pp-cli schedules list-v1` — List all schedules. You can also paginate the results.
- `trigger-dev-pp-cli schedules update-v1` — Update a schedule by its ID. This will only work on `IMPERATIVE` schedules that were created in the dashboard or...

**tasks** — Manage tasks

- `trigger-dev-pp-cli tasks` — Batch trigger tasks with up to 1,000 payloads with SDK 4.3.1+ (500 in prior versions).

**timezones** — Manage timezones

- `trigger-dev-pp-cli timezones` — Get all supported timezones that schedule tasks support.

**waitpoints** — Manage waitpoints

- `trigger-dev-pp-cli waitpoints complete-token-callback-v1` — Completes a waitpoint token using the pre-signed callback URL returned in the `url` field when the token was...
- `trigger-dev-pp-cli waitpoints