odd-studio 1.0.0 → 2.0.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "odd-studio",
-  "version": "1.0.0",
+  "version": "2.0.0",
   "description": "Outcome-Driven Development for Claude Code — a planning and build harness for domain experts building serious software with AI.",
   "keywords": [
     "claude-code",

package/skill/SKILL.md

@@ -75,6 +75,23 @@ Type `*plan` to continue planning, `*build` to enter build mode, or `*status` fo
 
 ---
 
+## MANDATORY RUFLO CHECKPOINTS
+
+These are non-negotiable tool executions. You MUST call these ruflo tools — not describe them, not summarise them. Actually invoke the tool at each gate.
+
+| Gate | Tool call required | When |
+|---|---|---|
+| Persona approved | `mcp__ruflo__memory_store` key `odd-persona-[name]` namespace `odd-project` | Immediately after Diana marks a persona approved |
+| Outcome approved | `mcp__ruflo__memory_store` key `odd-outcome-[name]` namespace `odd-project` | Immediately after Marcus marks an outcome approved |
+| Contract map complete | `mcp__ruflo__memory_store` key `odd-contract-map` namespace `odd-project` | Immediately after Theo completes the contract map |
+| Plan approved | `mcp__ruflo__memory_store` key `odd-plan` namespace `odd-project` | Immediately after Rachel's plan is approved |
+| State update (all stages) | Write updated `.odd/state.json` | After every persona, outcome, or plan approval |
+| Session start | `mcp__ruflo__memory_retrieve` key `odd-project-state` namespace `odd-project` | Before displaying any welcome or status message |
+
+**If ruflo is not available:** continue without it and note to the user that cross-session memory will not persist this session.
+
+---
+
 ## Core Commands
 
 All commands begin with `*`. When a user types a command, route to the correct behaviour below.
@@ -159,6 +176,22 @@ Initialise the ruflo swarm for parallel build execution. See the full Ruflo Swar
 
 ---
 
+### `*agent`
+
+Create a custom agent for a domain-specific concern. The domain expert describes what they need in plain language. ODD Studio configures the agent, assigns it to the swarm, and includes its reports in the verification output alongside the QA agent's report.
+
+Common use cases: brand voice (flag text that does not match the platform's tone), compliance (check every outcome against a regulatory standard), accessibility (review every screen against WCAG).
+
+The domain expert does not write agent code. They describe the concern:
+
+> "Every piece of text a customer sees should sound like it comes from a small, friendly independent bookshop. Flag anything that sounds corporate or uses jargon."
+
+ODD Studio creates the agent from that description. It runs on every relevant outcome and reports in domain language.
+
+After collecting the description, call `mcp__ruflo__agent_spawn` with the custom role and the domain expert's description as instructions. Confirm the agent is active and will run during the next `*build` or `*swarm` session.
+
+---
+
 ### `*export`
 
 Generate the IDE Session Brief. This is a standalone document that a developer or AI coding agent can use to execute a build session without needing to ask planning questions.
@@ -182,16 +215,24 @@ After writing the file, display: "Your Session Brief has been written to docs/se
 Load coaching content from chapter n of the ODD methodology book. Route to the appropriate file in `docs/chapters/`. If the chapter file does not exist, explain what that chapter covers conceptually and offer to explore it through dialogue.
 
 Chapter reference:
-- Chapter 1: Why features fail and outcomes succeed
-- Chapter 2: Persona depth — beyond demographics
-- Chapter 3: Writing outcomes that actually specify behaviour
-- Chapter 4: The walkthrough as specification
-- Chapter 5: Contracts — the grammar of system connections
-- Chapter 6: The Master Implementation Plan
-- Chapter 7: Build Protocol — how AI agents work from your plan
-- Chapter 8: Verification in the domain expert's language
-- Chapter 9: UI excellence for non-designers
-- Chapter 10: Handling change without breaking the plan
+- Chapter 1: It's All About Clarity
+- Chapter 2: The Right Division of Labour
+- Chapter 3: Features Aren't Enough
+- Chapter 4: Outcomes Should Be Specific
+- Chapter 5: Personas Are Load-Bearing
+- Chapter 6: Every Outcome Has a Contract
+- Chapter 7: Design the Outcome Twice
+- Chapter 8: The Master Implementation Plan
+- Chapter 9: Start from Zero
+- Chapter 10: The Build Protocol
+- Chapter 11: Verification Is Your Job
+- Chapter 12: Building One Outcome
+- Chapter 13: Concurrent Outcomes and the Swarm
+- Chapter 14: The Things That Scare You
+- Chapter 15: Good Interfaces Are Specified, Not Designed
+- Chapter 16: Managing Change
+- Chapter 17: The Swarm in Depth
+- Chapter 18: Conclusion
 
 ---
 
@@ -228,7 +269,8 @@ Display this reference:
 | `*contracts` | Map contracts with Theo |
 | `*phase-plan` | Build the Master Implementation Plan with Rachel |
 | `*ui` | Load UI excellence principles |
-| `*swarm` | Initialise ruflo swarm manually |
+| `*swarm` | Build all independent outcomes in the current phase simultaneously |
+| `*agent` | Create a custom agent for a domain-specific concern |
 | `*export` | Generate IDE Session Brief |
 | `*chapter [n]` | Load methodology coaching for chapter n |
 | `*why` | Explain why the current step matters |