docs: add Formula + Cook terminology and example formulas (gt-8tmz)

Updates to molecule-algebra.md:
- Added Formulas and Cooking section
- Four-tier model: Formula → cook → Proto → pour/wisp → Mol/Wisp
- Breaking Bad × Mad Max naming (Let's cook!)
- Updated vision diagram

New example formulas:
- shiny.formula.yaml (Engineer in a Box)
- rule-of-five.formula.yaml (Jeffrey's Rule expansion macro)
- security-audit.formula.yaml (AOP aspect)
- shiny-enterprise.formula.yaml (composed example)

.beads/formulas/shiny-enterprise.formula.yaml

@@ -0,0 +1,66 @@
+# shiny-enterprise.formula.yaml
+# Full enterprise engineering workflow with all the bells and whistles
+
+formula: shiny-enterprise
+extends: shiny
+description: |
+  Enterprise-grade engineering workflow.
+
+  Builds on Shiny with:
+  - Rule of Five on implementation (5x refinement)
+  - Security scanning (pre and post)
+  - Performance testing branch
+  - Review loop until approved
+  - Full logging
+
+  Cook this to get a ~30 step proto ready for serious work.
+version: 1
+
+compose:
+  # Apply Rule of Five to the implement step
+  # This expands implement → 5 steps: draft, refine-1..4
+  - expand:
+      target: implement
+      with: rule-of-five
+
+  # Apply security aspect to all implementation steps
+  # Adds security-prescan before and security-postscan after
+  - aspect:
+      pointcut: "implement.*"
+      with: security-audit
+
+  # Gate on security approval before submit
+  - gate:
+      before: submit
+      condition: "security-postscan.approved == true"
+      message: "Cannot submit without security approval"
+
+  # Parallel performance testing branch
+  # These run in parallel after implementation, before review
+  - branch:
+      from: implement.refine-4
+      steps:
+        - id: perf-test
+          description: Run performance benchmarks
+        - id: load-test
+          description: Run load/stress tests
+        - id: chaos-test
+          description: Run chaos engineering tests (fault injection)
+      join: review
+
+  # Loop review until approved (max 3 attempts)
+  - loop:
+      step: review
+      until: "review.output.approved == true"
+      max: 3
+      on-max: escalate
+
+  # Logging on all steps (AOP-style advice)
+  - advice:
+      target: "*"
+      before:
+        - id: log-start
+          description: "Log: Starting {step.id}"
+      after:
+        - id: log-end
+          description: "Log: Completed {step.id} with status {step.status}"