See
reality.md
A memory layer your agents can read before they act.
It buys you
agents stop guessing
Open method · working assessment
An open five-move method, local system-gap assessment, and exportable architecture brief for building dependable AI workflows.
Export preview
# Research loop — Architecture Brief First gap: Move 02 · Design Build next: one-job system specification Acceptance test: Another person can explain the workflow, owner, boundary, and stop condition. Guardrail: Do not automate an ambiguous job.
The complete export adds five scores, a seven-day build order, privacy note, and a decision to keep, revise, or stop the system.
Dependency map
See, Design, Build, Automate, and Compound are ordered. The assessment stops at the first gap so the recommendation remains buildable.
Method map
Next action
Find the first unlocked move
reality.md
A memory layer your agents can read before they act.
It buys you
agents stop guessing
one spec
A written contract for the repeating job you want handled.
It buys you
decisions stop leaking
one agent
A small named worker with one job and visible boundaries.
It buys you
first hour returns
one loop
A trigger, an agent, a review gate, a handoff, and a clear done state.
It buys you
repetition becomes supervised
one signal
A scoring layer that biases the next run toward what worked.
It buys you
the system learns
The moves are dependent. Do not automate before you can see. Do not build before the spec is clear. The assessment exists to find the first gap.
Run the assessmentProduct path
Assessment, architecture brief, method, standard, and starter templates.
Planned assessment pack; no checkout until the files, license, support, price, and refund terms are complete.
A scoped architecture review with availability and deliverables confirmed before payment.
Run the local assessment and export a Markdown architecture brief. No account or email required.