#Forum — 2026-07-07

When: Tuesday, July 7, 2026, 5–6 PM ET · Google Meet

#Ideas

Anything you want to talk about. Each idea gets its own heading, a sentence or two of context, and a Notes line we fill in when we discuss it.

#Weekly "bricks" — zrosenbauer

Each week you commit to one thing and deliver it as your primary focus (e.g. "move to a monorepo format with Turborepo"), and tie these into the weekly engineering updates for the org.

Notes:

#Release process + Changesets — zrosenbauer

Adopt Changesets for versioning and define the whole release flow — how do we release daily, possibly on every merge to main?

Notes:

#AI-authored PRs need a stated goal — zrosenbauer

For AI-authored PRs (Devin or similar), the PR needs to clearly explain what the goal was up front before anyone can review it.

Notes:

#Code standards — zrosenbauer

Define shared code standards for the org, living under docs/standards/**. Keep them terse and skimmable — massaman's AGENTS.md is a good shape to model, an Always / Never / Ask-First boundaries list.

Split agent instructions (AGENTS.md / CLAUDE.md) from the coding standards themselves — right now they're tangled. The standards should be their own docs so we can change them independently; agent instructions then just point at them.

Break each rule out on its own line so it's easy to grok. Starters:

• Named functions over arrow functions — reason TBD.
• kebab-case for everything on the filesystem — avoids the case-insensitivity mismatch between macOS and Linux.

Notes:

#Define our engineering process — the handbook ("Ways of Working"?) — zrosenbauer

The core problem for tomorrow. We don't have a defined process for how engineering plans, builds, ships, and QAs — end to end, idea → production. This is the foundation: it will drive every process decision across the engineering org, and product aligns to it. Goal of the forum is to get this going and agree the shape.

Scaffolded a light draft at docs/handbook/ — The Hitchhiker's Guide to Holly Engineering. Sections:

• Foundation — Overview, Values, Ways of Working
• Lifecycle (idea → shipped) — SDLC, Definition of Ready, Definition of Done, Release & QA
• Operations — Incident Response, On-Call
• Security — Secure Coding, Secrets & Access
• Architecture — how we decide + record (ADRs) and a light view of the system
• Craft — Engineering Standards, AI Development Standards

Reusable forms live in docs/templates/ (PRD, ADR, postmortem) and the decision records themselves in docs/adr/ — chapters link to them rather than embedding templates.

Name: going with "The Hitchhiker's Guide to Holly Engineering" (witty, memorable — the Don't Panic energy fits a reference you actually want people to open). "Ways of Working" is a chapter (team norms), not the book's name.

Notable calls (backed by a scan of real public handbooks — GitLab, PostHog, Trussworks, Obvious, Microsoft): incidents/on-call broke out into Operations; Security got its own section; no engineering-metrics chapter (lean handbooks don't — it's a dashboard concern); templates and ADR records moved out of the handbook so we don't scatter bare templates through the chapters.

Agenda (don't boil the ocean — lock the frame, not the content):

• 0:00–0:05 — Why: this is our operating model; product aligns to it.
• 0:05–0:20 — Ratify name + section structure. Timebox; decide the name today.
• 0:20–0:35 — Assign a chapter owner for each chapter (highest leverage).
• 0:35–0:50 — Draft one chapter live end-to-end as the template — DoD (small, high-agreement) — so everyone sees the shape and altitude.
• 0:50–0:58 — Agree the contribution model: lives in this repo, change by PR to the owner, "last reviewed" date per chapter.
• 0:58–1:00 — Homework: each owner drafts a one-paragraph stub before next forum.

Defer explicitly (owned + dated, not solved live): SDLC detail, AI standards specifics, incident sev definitions.

Notes:

#Notes

Overall takeaways and decisions — anything that didn't belong to a single idea above.