Named product
Guided Flow Editor
Working product on the shelf: author steps, branches, and handoffs as a first-class DAG, then repair broken graphs with the fix-flow CLI. Demo.
Failure mode → proof
Broken multi-step graphs → visible path + fix-flow
Failure mode: prompt/agent chains that fail mid-run with no inspectable graph. Scoped-fix proof: map the path as a DAG, guide the next step with AI, repair with fix-flow — days locked on a ≤3-day card before work starts. No WarBot outcomes claimed here.
How we work
Listen → card → fixed scope → ship
Name the failure mode. We write a one-page card with boundaries and days (≤3). You approve the card. We ship that scope — nothing adjacent slips in.
On the shelf
Guided Flow Editor
The named working product you can open today. Everything else is R&D or off.
Product
Demo
Guided Flow Editor
Visual DAG + AI guided flows + fix-flow CLI.
Author the path operators follow before runtime: steps, branches, and handoffs as a first-class artifact. Demo and iterate without inventing process mid-run — and repair broken multi-step graphs with the CLI when they fail.
Facial recognition product: OFF — not on the shelf.
R&D · not a product
Building WarBot
Android and emulator control with device stewardship for operators who need verified taps — a lords-bot.com-class niche. Status: coming / in development / not shipped. Not Live, not Open, not a buy-or-run SKU, and not warbot.cloud as shipped proof.
If your failure mode is blind device automation that taps without verifying screen state, say so in discovery. We may listen. We will not sell WarBot, treat warbot.cloud as shipped proof, or put a buy/run button on this page.
Delivery
Listen → card → fixed scope → ship
Home + CTA work stays inside a ≤3-day box once the card is approved.
1
Listen
You name the failure mode that keeps repeating.
2
Card
We write a one-page card: outcome, boundaries, ≤3 days.
3
Fixed scope
You approve the card. That is the whole job.
4
Ship
We deliver the card. No scope creep by default.
Discovery intake
Stuck on this failure mode?
Tell us the failure that keeps biting — Guided Flow graphs, broken multi-step agents, or a real device-stewardship pain. Not a blank wishlist. We reply with a card (or a clear no) before anything is scoped to ship.