Agents Playbook

The matrix — pillars × phases

This is the master content map. Each cell points to the practices that apply when **pillar** meets **SDLC phase**.

The matrix — pillars × phases

This is the master content map. Each cell points to the practices that apply when pillar meets SDLC phase.

Cells marked (stub) are scaffold-only; content lands in subsequent sessions. Cells marked are shipped.

01 Discover02 Design03 Build04 Test05 Ship06 Operate
ArchitectureDefine mental map; pick 5–7 logical groupsADR before structure change; RFC before breaking contractModular boundaries; named exports; Zod at every edge; sized filesContract tests; cross-package import lintsVersioned releases; peer-dep compat matrixTrack tech debt against ADRs
SecurityThreat model; data classificationVault scope; RBAC roles; consent vs elevationSealed secrets; egress allowlist; signed audit ledgerSecurity review per PR; pen-test gatesSigned artifacts; key rotation runbookIncident response; break-glass audit
UI-UXAudience taxonomy; surface inventoryDesign tokens; primitives catalog; motion budgetNo raw \<button\>/\<input\>; intl on every string; useT() everywhereVisual regression; a11y screen-reader passBrand-kit/whitelabel build matrixA/B + delight loops; empty-state honesty
QualityDefine "done" per surface (completeness contract)Test pyramid; coverage targets per packageFile-size budgets; lint rules; hermetic testscheck:all, sanity, structural gates, e2ePre-push hooks; release gatesBug-hunt cadence; mutation testing
GovernanceDecision log culture; ADR/RFC processesPR intent manifest schema; merge rulesRemoves-list discipline; concurrent-agent awarenessMulti-agent review pipelineChangesets; semver disciplinePostmortems; tombstone retired plans
AI-collaborationCLAUDE.md, AGENTS.md, MEMORY.md bootstrapSlash commands; sub-agent recipes; system promptsGoal-mode loop; one-sub-unit-per-session ruleVerify-first close; duplication-claims-API-groundedPhased PR + admin mergePersistent memory; lessons graph

Reading order

If you adopt left-to-right (by phase), agents can ramp incrementally. If you adopt top-to-bottom (by pillar), you can roll out one concern across the whole SDLC.

Adoption modeStart atThen
Greenfield projectphases/01-discover/templates/CLAUDE.md.template.mdpillars/architecture/universal.md
Brownfield retrofitpillars/quality/README.md (gates first)pillars/governance/README.md (PR intent + merge rules)
Just need agent rulestemplates/CLAUDE.md.template.mdpillars/ai-collaboration/README.md
Just need design systempillars/ui-ux/README.mdtemplates/ design-tokens recipe

Status legend

  • ✓ Shipped (read it)
  • ◐ Scaffolded with scope; content partial
  • (stub) README placeholder only; no body content yet

Current status (v0)

UniversalTS-concrete
architecture
security
ui-ux
quality
governance
ai-collaboration
PhaseStatus
01 discover
02 design
03 build
04 test
05 ship
06 operate
TemplatesStatus
ADR / RFC
PR intent
CLAUDE.md / AGENTS.md / MEMORY.md
PromptsStatus
System (architect, implementer, reviewer, security)
Sub-agent recipes (explore, plan, code-explorer, code-reviewer)
Slash commands (goal, loop, review, clear, sanity, ship)
OtherStatus
Scripts (gate reference impls)✓ all 12 gates + orchestrator
Phases (deep content)