| name | build-discipline |
| description | Defines the shared Building loop and named sub-loop entry points so dashboards, wikis, and MCPs each follow a consistent but distinct flow. Use when the user says "build discipline", "how do I build X", "start a new dashboard", "start a new wiki", "start a new MCP", "start a new agent", "build loop", or "which sub-loop". Also triggers when the user is about to kick off any building work and needs to pick a track.
|
Boot from os/PREAMBLE.md.
Build Discipline
The universal Building loop for all Gal-owned work. Pick the right sub-loop, follow it, don't shortcut.
The Loop
discover → spec (to-prd) → prototype → build → attest → ship
Every build — no matter how small — touches all six stages. Size decides depth, not whether you do them.
| Stage | What happens | Key skill |
|---|
| discover | Clarify the question, identify the right sub-loop, check what already exists | kb-query, github-repo-analyzer |
| spec | Formalise intent as a testable PRD or one-pager | to-prd |
| prototype | Answer one design question with throwaway code/query | prototype |
| build | Implement against the spec | sub-loop specific (see below) |
| attest | Verify the output matches the spec; run the relevant review | data-review (data), git-guardrails (code) |
| ship | Publish / deploy / commit with context pointer | commit-push-pr, deploy-mf, n8n deploy |
Named Sub-Loops
data / monitoring
Entry: You have a question that needs a number. Start at prototype (write a scratch query first).
prototype → query → viz → deliver
| Step | Tool / Skill |
|---|
| query | Kremer MCP (kremer-mcp), z2h (z2h-explore), Snowflake direct |
| viz | monday-data-viz-vibe |
| deliver | to-prd (for spec), monday-presentation-v2 or slack-in-my-voice (for comms) |
| pre-ship gate | Run data-review before sharing anything externally. |
knowledge
Entry: You need to add, query, or compile governed knowledge. The engine is llm-wiki-agent.
ingest → compile → query / recall / research
| Command | Purpose |
|---|
/wiki-init | Bootstrap a new wiki in a repo |
/wiki-compile | Re-build the wiki index from sources |
/learn-capture | Add a document or URL as a source |
wiki-query / learn-recall | Answer questions from the compiled wiki |
learn-research | Deep research that files back into the wiki |
/add-tool <doc> | doc → doc-to-skill → commit (adds an executable skill from a reference doc) |
MCP / agent
Entry: You need a new autonomous capability — either an agent or an MCP server.
design → validate → build → register → test → ship
| Step | What |
|---|
| design | open_claw agent design (roles, tools, guard rails) |
| validate | nanoclaw-validate (structure check before writing code) |
| MCP build | GTM Gateway pattern; read skills/packs/gtm/ for conventions |
| test | Dry-run in staging before prod deploy |
| ship | Register in os/AGENTS.md; commit with context pointer |
OS-vs-Team-Repo Boundary
This matters. Duplication here wastes hours.
| Layer | What lives here | Examples |
|---|
OS (sir_albert) | Gal's reusable skills — logic Gal owns and reuses across contexts | to-prd, data-review, build-discipline |
| Team repos | Shared / governed knowledge and team-specific workflows | marketing-cookbook (GTM playbooks), mopa_brain (MOPA logic), data-cookbook (data definitions) |
Rule: If it's a reusable capability → OS. If it's governed team knowledge → team repo. When in doubt, ask: "Would another team member need to maintain this?" If yes → team repo.
Guardrails
- Never skip
attest. A shipped thing that wasn't attested is a liability.
data-review is mandatory before any dashboard or analysis goes external.
- The loop applies to one-liners too — it's just faster, not absent.