A your AI Code harness for building apps in TypeScript monorepos: intake, PRD, a generated skeleton with CI, backlog, independent verification and rules backed by checks.
Skills 14 in the Claude listing
Build story Build one backlog story end to end with independent checks — done-contract from its dimensions and invariants, acceptance tests written by the verifier before any code, implementation, a cold review a Check bugbash Full bug bash of a Keelokit project across every dimension (data, API, contracts, logic, integrity, auth, UX, UI, web, mobile, i18n, copy, a11y, security, NFRs, config, ops, and dx, docs and packaging Check health Check and repair the health of a Keelokit project's harness — rules without enforcers, enforcers that point nowhere, expired or incomplete exceptions, gaps in docs/context, malformed stories — and add Check security Security and privacy review of a Keelokit project, in depth — a map of every piece of personal data (what, why, where, who sees it, how long), the privacy law of each market, a threat model of the cri Harness upgrade keelokit Entry point of the Keelokit harness. Plan backlog Turn a product's context and PRD into a backlog — epics, stories as vertical slices, consolidated, prioritised and grouped into waves that never touch the same files. Plan intake Interview the user in their language (and read what already exists) to write a product's context in docs/context — problem, users, domain, constraints, environments, and every open gap with owner and Project adopt Bring an existing repository (brownfield) under the Keelokit harness without touching its code — installs only .keelokit/, maps each house rule to the checks the repo already has, registers what it ca Project dashboard Open or refresh the project's Keelokit dashboard — one short, branded page for day-to-day work, kept live — where the project is, the next step with its command to copy, what waits for you (approvals, Project new Start a new product from zero with the Keelokit harness — intake interview, PRD, stack decision, monorepo skeleton with CI and staging from day 1, and the first backlog. Project report Generate the project's complete status report — a read-only Keelokit page with every stage and its documents (context, gaps, PRD scope and metrics, stack, backlog by wave and by epic), bug bashes and Ship release Release a version of the product — pick the next semver from what landed since the last tag, write the CHANGELOG entry in words users understand (stories done, bug fixes, what changed for them), tag v Ship setup Set up where the product runs — staging and production — for people who have never deployed anything.