Outcome chain
Business outcome to operations outcome to requirement to build item, each with acceptance criteria and how it is measured.
Works with GitHub Spec Kit and AWS Kiro
Connect a GitHub repository. OT reads the specs already in it and drafts the outcomes, operations pages, registers and readiness report your stakeholders need. Your team confirms it, and every save is a pull request.
Read only until you choose to save. Nothing reaches your default branch without a merged pull request.
How it works
Sign in with GitHub and pick a repository. OT finds the Spec Kit or Kiro specs, tasks, commits and pull requests that are already there.
Claude works backwards from the specs to business outcomes, operations pages, registers and a readiness report. Everything it proposes is marked inferred until a person confirms it.
Edit and confirm in OT. Every save opens a pull request that updates .ot/ and patches the specs, so developers keep building from the current plan. Merging is the approval.
A GitHub Action keeps OT current on every merge.
What you get
Business outcome to operations outcome to requirement to build item, each with acceptance criteria and how it is measured.
One clear operating page per outcome: who runs it, how it is checked and when it is next reviewed.
Risks, decisions, changes, benefits and delegations, seeded from your specs and kept in step with every merge.
Traceability metrics and a plain list of what is missing before you go live.
Principles
Extracted from a spec, inferred by Claude, drafted in OT, confirmed by a person or retired. Claude never overwrites what a person has confirmed.
OT lives in a .ot folder in your repository. The app keeps only accounts and unsaved drafts. Client data stays with you.
OT never pushes to your default branch. Changes arrive as labelled pull requests, and the merge is recorded in the change register.
Scans are deterministic, so pull request diffs show only real change. Specs are patched line by line, never rewritten.