Forge 42 — Build Together
Ship working software every two weeks—without burning the roadmap. Forge 42 is our disciplined engineering phase: code generates where it should, humans craft where it counts, and every commit passes the same quality gates we'll use in production.
1 · Why Forge 42?
| Pain we've seen | How Forge 42 fixes it |
|---|---|
| 80 % of a sprint lost to boiler-plate | Code-gen CLI scaffolds UI, API, tests in minutes. |
| "Done" means different things to dev vs. QA | Definition-of-Done inherits BDD scenarios from Lab 42—green tests or it doesn't merge. |
| Late surprises in performance or security | Static analysis, k6 load, OWASP scans run on every PR; fail fast, fix cheap. |
| Stakeholders can't see progress | Deploy previews: click a link, play with the increment in your browser. |
3 · Toolchain & Accelerators
| Layer | Tools & Footnotes |
|---|---|
| UI | SvelteKit + Tailwind + Storybook component library (tables, wizards, charts). |
| API | Django DRF, GraphQL (Strawberry), pgvector for AI search. |
| DevOps | GitHub Actions, Docker, Fly.io automations; scratch-org scripts for Salesforce work. |
| Testing | PyTest & Coverage • Jest • Cypress • k6 • OWASP ZAP. |
| Quality Gates | ESLint, Bandit, Snyk, CodeQL—block merge if threshold missed. |
| Observability | Sentry, OpenTelemetry, Datadog dashboards wired from day 1. |
4 · Cadence & Ceremonies
| Cadence | Event | Outputs |
|---|---|---|
| Daily (15 min) | Stand-up | Blockers, plan, burn-down pulse |
| Each PR | Code review + automated gates | Green check = deploy preview |
| End of Sprint (2 wks) | Review & Retro | Demo URL, velocity stats, improvement actions |
| Monthly | Steering Day | Budget burn, FCI delta, risk heat-map |
All events are open-door or recorded for async viewing.
5 · Quality Gates (merge-blockers)
| Category | Threshold |
|---|---|
| Unit & BDD tests | ≥ 85 % line coverage |
| Static analysis | 0 high-severity issues |
| Performance | API p95 ≤ 300 ms (target) |
| Security | No critical CVEs; OWASP ZAP score "pass" |
| Accessibility | WCAG 2.1 AA for primary flows |
Fail a gate → PR auto-blocks → engineer fixes before merge.
6 · Risk Controls
- Red-Flag Register updated each sprint.
- Error budgets (SRE) tie performance regressions to deployment pausing.
- Feature flags for safe rollout / revert without hot-fix branches.
7 · Sample 3-Sprint Timeline (8-week MVP)
| Sprint | Theme | Key Deliverables |
|---|---|---|
| 1 | Core scaffold & auth | Repo, CI/CD, login, smoke tests green |
| 2 | CRUD + UI skeleton | Entity forms, table view, Cypress story |
| 3 | Integrations + perf harden | API mesh, p95 < 300 ms, demo to pilot users |
At Sprint 3 we re-score FCI; if ≥ 8, you can freeze scope into Fixed Pack or keep cruising in Pod/DaaS.
8 · Mini Case
BIBC WhatsApp Engine — Forge 42 built & shipped a Svelte frontend, Django API and pgvector search in 4 weeks. Deployed through GitHub Actions to Fly.io; first production spike handled 300 k users / day at p99 < 350 ms.
9 · FAQs
10 · Call to Action
Ready to see real software by next sprint review?
Book a Forge 42 ignition call—get squad lineup, first sprint goal and start date in 48 hours.
Kick-off Forge 42