> workflows · Step 3

Lifecycle workflows

After onboarding and installing Your kit — pick the stage that answers what should I do next? Each page links to real corpus artifacts (prompts, skills, patterns, gates).

  1. 01

    Idea

    Turn a raw product or engineering idea into a bounded initiative with explicit non-goals, success metrics, and an initial risk tier before any repo work starts.

  2. 02

    Project Kickoff

    Align repo, context package, hooks, and model/tool policy so the team has a Definition of Ready before architecture deepens.

  3. 03

    Architecture

    Produce ADRs, interface contracts, and threat-model sketches that bound what AI agents may invent during implementation.

  4. 04

    Implementation Planning

    Break approved architecture into verifiable slices with tests, rollback, and evidence hooks — before any feature coding begins.

  5. 05

    Human Review

    Approve the implementation plan (and risk posture) before any AI coding — a plan-approval HITL gate, not a pull-request review.

  6. 06

    AI Coding

    Implement only approved plan slices inside hooks, stack skills, and architecture boundaries — with tests and docs landing in the same change.

  7. 07

    Code Review

    Human-accountable PR review of AI-assisted changes for correctness, security, and maintainability — AI assists; it does not approve.

  8. 08

    Testing

    Prove behavior with automated tests and eval gates before promoting a release candidate — especially for AI-touched paths.

  9. 09

    Security

    Validate threat model, dependencies, secrets, and AI-specific attack surfaces before performance tuning or release plumbing.

  10. 10

    Performance

    Meet latency, cost, and throughput budgets for application and AI paths with profile-guided changes — not speculative rewrites.

  11. 11

    DevOps

    Automate build, deploy, and environment promotion with immutable artifacts, secrets hygiene, and safe progressive delivery defaults.

  12. 12

    Monitoring

    Observe production behavior, model quality drift, and incident signals with actionable alerts — not vanity dashboards.

  13. 13

    Production

    Operate the live release with go/no-go discipline, progressive exposure, verified rollback, and customer communication readiness.

  14. 14

    Continuous Improvement

    Convert incidents, retros, and metrics into updated prompts, skills, evals, and debt paydown — then feed the next idea cycle.

Learn · Operate · Assure

Systematic surfaces after the harness is installed — not a raw folder dump.