Fob
Browse the docs

Use with AI agents

Coding agents such as Claude Code, Cursor or Codex can write a step, run it with fob-worker, read the error, fix it and run it again. That loop is the same one you'd follow by hand, and fob-worker keeps it on your machine.

Video coming soon
Write and debug a worker step with Claude Code and fob-worker (3–4 min)

1. Set up the repo yourself#

Create the worker repo, fill in .env and pull your station files (Get started) before you hand it to an agent, so the agent never needs your API key.

2. Tell the agent about the repo#

The template's CLAUDE.md already describes the layout, defineStep() and the station files. Add the fob-worker loop to it (or to AGENTS.md):

## Running and debugging steps

- List steps: `fob-worker steps list`. Run one: `fob-worker steps run <slug> --scenario <name>` (or `--station <CODE>`, or `--empty`).
- Save test configs as `.orchestrator/scenarios/<slug>/<name>.json`. Outputs land in `temp/<slug>.json` and feed the next step's `step_outputs`.
- On failure, read the `Resolved config` line and the error; rerun with `DEBUG=1` for a stack trace.
- Workpieces: `fob-worker lines status <LINE>`, `fob-worker workpieces show <id>`, and `error.txt` in the failed bin.
- Never push station files (`fob-orc stations push`) or run stations on the Orchestrator without asking me.
- Docs: https://finopsbricks.com/docs/workers (raw Markdown: https://finopsbricks.com/llms.txt)

3. Ask#

  • "Add a step IN1_02_report that attaches a markdown report of the flagged invoices. Write a scenario and run it."
  • "IN1_01_check fails with the two-invoices scenario. Find out why and fix it."
  • "INV-1002 is stuck in IN1/failed. Read its error and tell me what happened."
  • "Write Jest tests for the parsing function in IN0_01_seed."

Docs for agents#

Every docs page is also served as raw Markdown: