"Use workers for Unstract's Celery and Postgres-backed queue
Scanned 9/9/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill workers --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Workers?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-workers)More formats (shields.io, HTML) on the badges page.
---
name: "workers"
description: "Use workers for Unstract's Celery and Postgres-backed queue
workers, worker launchers, queue routing, health ports, and operations."
disable-model-invocation: true
metadata:
disco-role: operating
license: AGPL 3.0
---
# Workers
Use this sub-skill when the task is about asynchronous processing in Unstract: worker launchers, queue selection, Celery / PG-queue routing, worker health, logging, or operations and recovery.
## Owns
- `workers/run-worker.sh`, `workers/run-worker-docker.sh`, `workers/worker.py`, and the worker subpackages under `workers/`.
- Celery workers such as `api-deployment`, `general`, `file_processing`, `callback`, `notification`, `scheduler`, `log_consumer`, `executor`, and `ide_callback`.
- PG-queue workers: the consumer, the reaper, and the named PG queue roles.
- Worker tests, worker health ports, queue routing, metrics, and operational guidance.
## Excludes
- Backend route families, auth, and hosted MCP internals — use `backend-platform`.
- Full-stack bootstrap and container orchestration — use `platform-deployment`.
- Shared SDK / tool packages and tool protocol docs — use `sdk-and-tools`.
- Frontend runtime config or route changes — use `frontend`.
- Repo test-rig manifests and coverage reporting — use `testing-rig`.
## Start Here
Read `references/worker-matrix.md` first when you need to understand:
- which worker type owns a queue,
- which health port it uses,
- how the Celery set differs from the PG-queue set,
- or how the launcher CLI maps aliases to worker directories.
Read `references/troubleshooting.md` when the issue is a worker that does not start, drains the wrong queue, or fails a health / callback check.
For quick inspection, the usual first command is the launcher help path:
```bash
cd workers
./run-worker.sh --help
```
## Shared References
- `references/worker-matrix.md` — worker aliases, queues, ports, and launch modes.
- `references/troubleshooting.md` — worker startup, queue routing, health, and PG-queue issues.
- `../references/service-map.md` — repo-wide service map and dependencies.
- `../references/installation-and-env.md` — install and environment matrix for worker workflows.
- `../references/repo-provenance.md` — source snapshot used to build this skill.
## Common Task Routing
| User request | Read next |
| --- | --- |
| "Which worker handles this queue?" | `references/worker-matrix.md` |
| "How do I start a specific worker?" | `references/worker-matrix.md` |
| "How does the PG queue consumer or reaper work?" | `references/worker-matrix.md` |
| "Why is a worker not processing tasks?" | `references/troubleshooting.md` |
| "Why is health or metrics failing?" | `references/troubleshooting.md` |
## Safety Boundaries
- Do not assume live RabbitMQ, Redis, or PostgreSQL are available unless the user explicitly wants runtime validation.
- Do not confuse the Celery fleet with the PG-queue fleet; they are intentionally separate launch sets.
- Do not debug queue routing from the wrong worker directory; the launcher aliases and worker packages have their own ownership rules.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!