haiku-method 2.2.0

package/studios/software/templates/bugfix.md

@@ -0,0 +1,35 @@
+---
+name: bugfix
+studio: software
+description: Fix a reported bug with investigation, fix, and verification
+parameters:
+  - name: bug
+    description: Description of the bug or link to the bug report
+    required: true
+  - name: severity
+    description: "Bug severity: critical, high, medium, low"
+    required: false
+    default: medium
+stages-override: [inception, development]
+units:
+  - name: "investigate"
+    stage: inception
+    criteria:
+      - "Root cause identified with evidence (logs, reproduction steps)"
+      - "Fix approach documented"
+      - "Blast radius assessed (what else might be affected)"
+  - name: "fix"
+    stage: development
+    criteria:
+      - "Bug no longer reproducible"
+      - "Regression test written that fails without the fix"
+      - "No other tests broken"
+---
+
+Streamlined template for bugfixes — skips design and product stages.
+
+## When to Use
+
+- Reported bug with a clear reproduction path
+- Bug that doesn't require design changes or new product specs
+- Urgent fixes (set severity parameter to control priority)

package/studios/software/templates/new-feature.md

@@ -0,0 +1,62 @@
+---
+name: new-feature
+studio: software
+description: Build a new user-facing feature from requirements through deployment
+parameters:
+  - name: feature
+    description: What feature to build
+    required: true
+  - name: requirements_url
+    description: Link to requirements doc, PRD, or ticket
+    required: false
+  - name: design_url
+    description: Link to design file (Figma, etc.)
+    required: false
+units:
+  - name: "discovery"
+    stage: inception
+    criteria:
+      - "Domain model maps all entities for {{ feature }}"
+      - "Technical constraints documented (APIs, dependencies, performance)"
+      - "Unit elaboration covers the full feature with no gaps"
+  - name: "visual-design"
+    stage: design
+    criteria:
+      - "Screen layouts for all breakpoints (mobile, tablet, desktop)"
+      - "Interactive states defined (default, hover, focus, error, loading)"
+      - "Design tokens reference the existing design system"
+  - name: "behavioral-spec"
+    stage: product
+    criteria:
+      - "Happy path and error scenarios documented for each user flow"
+      - "Data contracts define request/response schemas"
+      - "Acceptance criteria are testable (Given/When/Then)"
+  - name: "implementation"
+    stage: development
+    criteria:
+      - "All acceptance criteria passing"
+      - "Test coverage >= 80% for new code"
+      - "No TypeScript any types without justification"
+  - name: "deploy"
+    stage: operations
+    criteria:
+      - "Health checks verify the new feature is live"
+      - "Monitoring covers the new endpoints/flows"
+---
+
+Standard template for building a new user-facing feature through the full software lifecycle.
+
+## When to Use
+
+- New feature with UI and API components
+- Feature with requirements/designs already defined (or to be defined during inception)
+- Work that needs the full inception → design → product → development → operations flow
+
+## Context Sources
+
+{{ if requirements_url }}
+- Requirements: {{ requirements_url }}
+{{ end }}
+{{ if design_url }}
+- Design: {{ design_url }}
+{{ end }}

package/studios/software/templates/refactor.md

@@ -0,0 +1,28 @@
+---
+name: refactor
+studio: software
+description: Restructure code to improve quality without changing behavior
+parameters:
+  - name: target
+    description: What to refactor (module, pattern, dependency)
+    required: true
+  - name: motivation
+    description: Why this refactor is needed
+    required: false
+stages-override: [inception, development]
+units:
+  - name: "assess"
+    stage: inception
+    criteria:
+      - "Current state documented (what exists, why it's problematic)"
+      - "Target state documented (what it should look like)"
+      - "Migration path identified (incremental steps, not big bang)"
+  - name: "refactor"
+    stage: development
+    criteria:
+      - "All existing tests pass without modification"
+      - "No behavioral changes (same inputs → same outputs)"
+      - "Code review confirms improved structure"
+---
+
+Restructure without behavior change — skips design and product.

package/studios/training/STUDIO.md

@@ -0,0 +1,11 @@
+---
+name: training
+description: Training program design and delivery lifecycle
+stages: [needs-analysis, design, develop, deliver, evaluate]
+category: back-office
+default_model: sonnet
+---
+
+# Training Studio
+
+Training program lifecycle from needs assessment through effectiveness evaluation. Covers skills gap analysis, curriculum design, content development, facilitation, and learning outcome measurement. Works for technical training, leadership development, onboarding programs, and compliance training.

package/studios/training/examples/new-hire-engineering-onboarding/intent.md

@@ -0,0 +1,18 @@
+---
+title: new-hire-engineering-onboarding
+studio: training
+stages: [needs-analysis, design, develop, deliver, evaluate]
+status: complete
+---
+
+# Engineering New-Hire Onboarding (30-day)
+
+Structured 30-day onboarding program for new engineering hires. Week one is tooling setup and pair programming; week two ships a real change to production; weeks three and four ramp up ownership of a discrete area. Explicit milestones and a graduation review at day 30.
+
+## Goals
+
+- Needs analysis across current engineering hires to identify onboarding gaps
+- Week-by-week milestones with clear "done" criteria
+- Week-two real ship target (every new hire ships something real by day 10)
+- Day-30 graduation review with manager + buddy + new hire
+- Measured: time-to-first-ship and time-to-first-on-call rotation

package/studios/training/examples/new-hire-engineering-onboarding/pre-intent.yml

@@ -0,0 +1,17 @@
+related:
+  - documentation
+  - hr
+  - project-management
+turns:
+  - role: user
+    text: "Build a new-hire onboarding program for the engineering team."
+  - role: agent
+    text: "How long is the program — week one, first month, first 90 days?"
+  - role: user
+    text: "First 30 days. Structured, with explicit milestones."
+  - role: agent
+    text: "Shipping a real change in week one, or ramp-up only?"
+  - role: user
+    text: "Shipping something real by end of week two. Week one is tooling and pairing."
+  - role: agent
+    text: "Got it — 30-day program, week one tooling + pairing, week two real ship, weeks three and four ownership ramp."

package/studios/training/fix-hats/reconciler.md

@@ -0,0 +1,10 @@
+---
+agent_type: general-purpose
+---
+**Focus:** Reconcile cross-stage artifacts against studio-wide standards. You are NOT wearing a stage-specific hat — you are resolving a whole-intent finding that spans stages. Your mandate is alignment, not fresh authoring.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** create new units, new scope, or new features
+- The agent **MUST NOT** modify unit workflow engine fields
+- The agent **MUST NOT** touch artifacts unrelated to the named finding
+- The agent **MUST NOT** re-open decisions settled at each stage's review gate

package/studios/training/fix-hats/validator.md

@@ -0,0 +1,10 @@
+---
+agent_type: general-purpose
+model: haiku
+---
+**Focus:** Independently verify that the reconciler's fix resolves the intent-scope feedback. You are the terminal hat — the workflow engine trusts your closure decision.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** edit any file — you verify, you do not fix
+- The agent **MUST NOT** close a finding that isn't actually resolved
+- The agent **MUST NOT** reject a finding because "it's not worth fixing" — either close, leave open, or reject as genuinely invalid

package/studios/training/review-agents/cross-stage-consistency.md

@@ -0,0 +1,17 @@
+---
+scope: intent
+interpretation: lens
+---
+**Mandate:** Verify the intent's artifacts are internally consistent across stages. You are the only reviewer that sees the whole intent at once — your job is to catch seams that per-stage reviewers miss.
+
+**Check:**
+- The agent **MUST** verify that each stage's outputs align with what upstream stages specified — no dropped requirements, no silent scope expansion
+- The agent **MUST** verify that naming is consistent across stages — a concept named one thing upstream should carry the same name downstream
+- The agent **MUST** verify that stages' declared outputs exist at the paths their unit frontmatter promised
+- The agent **MUST** verify that the stages collectively deliver the intent's stated goal (read `intent.md`) — partial delivery is a finding
+- The agent **MUST** verify that concerns raised by any stage's review were actually addressed (not silently ignored)
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** re-litigate decisions already approved at each stage's gate
+- The agent **MUST NOT** propose new features or scope additions
+- The agent **MUST NOT** flag stylistic preferences — concrete divergence only

package/studios/training/stages/deliver/STAGE.md

@@ -0,0 +1,17 @@
+---
+name: deliver
+description: Facilitate training delivery and coordinate logistics
+hats: [facilitator, coordinator, verifier]
+fix_hats: [facilitator, feedback-assessor]
+review: auto
+elaboration: autonomous
+inputs:
+  - stage: develop
+    output: training-materials
+  - stage: design
+    discovery: curriculum-plan
+---
+
+# Deliver
+
+Facilitate training delivery and coordinate logistics.

package/studios/training/stages/deliver/discovery/DELIVERY-LOG.md

@@ -0,0 +1,29 @@
+---
+name: delivery-log
+location: .haiku/intents/{intent-slug}/knowledge/DELIVERY-LOG.md
+scope: intent
+format: text
+required: true
+---
+
+# Delivery Log
+
+Session records, attendance data, and facilitator observations from training delivery.
+
+## Content Guide
+
+Structure the log for evaluation and improvement:
+
+- **Session records** -- dates, locations/platforms, facilitators, and session-specific notes
+- **Attendance data** -- participation and completion rates by session and module
+- **Facilitator observations** -- engagement patterns, areas of confusion, and content improvement suggestions
+- **Learner feedback** -- real-time feedback captured during sessions
+- **Logistics issues** -- problems encountered with resolution and prevention recommendations
+- **Curriculum deviations** -- any departures from the planned curriculum with rationale
+
+## Quality Signals
+
+- All planned sessions are accounted for with attendance data
+- Facilitator observations include specific, actionable content improvement suggestions
+- Issues are documented with both resolution and prevention strategies
+- Delivery data is sufficient to support the evaluation stage

package/studios/training/stages/deliver/hats/coordinator.md

@@ -0,0 +1,13 @@
+**Focus:** Manage delivery logistics to ensure sessions run smoothly and all participants have what they need.
+
+**Responsibilities:**
+- Coordinate scheduling, room/platform setup, and technical requirements
+- Distribute materials and manage access provisioning before each session
+- Track attendance and completion records
+- Resolve logistics issues rapidly to minimize disruption to learning
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST** verify technical setup before session start
+- The agent **MUST NOT** distribut materials too late for participants to prepare
+- The agent **MUST** track attendance systematically across sessions
+- The agent **MUST NOT** fail to have contingency plans for common logistics failures

package/studios/training/stages/deliver/hats/facilitator.md

@@ -0,0 +1,13 @@
+**Focus:** Deliver training sessions effectively and capture learner engagement data and content improvement opportunities.
+
+**Responsibilities:**
+- Facilitate sessions according to the curriculum plan and facilitator guide
+- Adapt delivery in real time based on learner engagement and comprehension
+- Document questions, areas of confusion, and suggested content improvements
+- Ensure all learners have opportunity to participate and practice
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** readd from materials rather than facilitating interactive learning
+- The agent **MUST NOT** ignore learner feedback signals during sessions
+- The agent **MUST** document observations that could improve future deliveries
+- The agent **MUST NOT** allow a few participants to dominate while others disengage

package/studios/training/stages/deliver/hats/feedback-assessor.md

@@ -0,0 +1,11 @@
+---
+agent_type: general-purpose
+model: haiku
+---
+**Focus:** Independently verify that a fix addresses the feedback finding as written. You are the terminal hat in this stage's fix-hat sequence — the workflow engine trusts your closure decision.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** edit any file — you are a verifier, not a fixer
+- The agent **MUST NOT** close a finding that isn't actually resolved — that is how drift hides
+- The agent **MUST NOT** reject a finding because "it's not worth fixing" — that is the human's decision, not yours; either close when resolved, leave open when not, or reject when genuinely invalid
+- The agent **MUST NOT** expand the scope beyond the one feedback item you were dispatched against

package/studios/training/stages/deliver/hats/verifier.md

@@ -0,0 +1,26 @@
+**Focus:** Validate the per-unit operational artifact for the deliver stage of training. Units here are delivery session — operational steps with concrete preconditions, actions, and post-condition checks. Validation rules check that preconditions are stated, the action is unambiguous, the post-condition has a verifiable check, and rollback is named where applicable.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** read or interpret unit frontmatter for any mechanical purpose. workflow engine territory per architecture §1.1.
+- The agent **MUST NOT** validate against frontmatter schema, `depends_on:` resolution, status-field shape, or any other FM-driven check — those are workflow engine responsibilities.
+- The agent **MUST NOT** advance a unit whose body is a placeholder, contains TODO markers, or has empty sections.
+- The agent **MUST NOT** reject for stylistic preferences. Substantive gaps only.
+- The agent **MUST** name a specific failed criterion in any rejection.
+- The agent **MUST NOT** invent rules not in this mandate. Stage scope is the contract.
+
+## What you check (BODY ONLY)
+
+### 1. Preconditions, action, post-condition all stated
+The unit body MUST have three concrete sections: preconditions (what must be true before the action runs), the action itself (one unambiguous procedure), and post-condition checks (how to confirm the action succeeded). Reject if any of the three is missing or vague.
+
+### 2. Verifiable post-condition
+The post-condition section MUST name a check that produces a clear pass/fail signal — a metric to read, a query to run, a screen to inspect with named expected values. "Verify by eye that things look good" is a reject.
+
+### 3. Rollback / recovery named where applicable
+Operational units MUST declare a rollback procedure OR explicitly state "no rollback — forward-fix only" with a rationale. Silent absence of rollback is a reject for any unit whose action is not idempotent.
+
+### 4. Decision-register consistency
+The unit must not propose an operational approach contradicting a recorded Decision (e.g., blue-green deploy when Decision N chose canary). Cite the Decision ID.
+
+### 5. Open questions accounted for
+Every "Open Questions" entry must be answered, defaulted, OR flagged `(needs human escalation)`. Operational open questions left to runtime are how outages happen.

package/studios/training/stages/deliver/outputs/DELIVERY-LOG.md

@@ -0,0 +1,25 @@
+---
+name: delivery-log
+location: .haiku/intents/{intent-slug}/stages/deliver/artifacts/
+scope: intent
+format: text
+required: true
+---
+
+# Delivery Log
+
+Training session records with attendance, feedback, and facilitation notes.
+
+## Expected Artifacts
+
+- **Session records** -- attendance, completion rates, and real-time learner feedback per session
+- **Facilitation notes** -- questions asked, areas of confusion, and suggested content improvements
+- **Logistics confirmation** -- technical setup, access provisioning, and material distribution verified
+- **Learner feedback** -- collected per session with trends identified
+
+## Quality Signals
+
+- Attendance and completion rates are recorded for every session
+- Facilitation notes capture areas of confusion for curriculum improvement
+- All technical setup is confirmed complete before each session
+- Feedback is collected and synthesized across sessions

package/studios/training/stages/deliver/phases/ELABORATION.md

@@ -0,0 +1,16 @@
+# Deliver Stage — Elaboration
+
+## Criteria Guidance
+
+### Good criteria — concrete and verifiable
+
+- "Delivery log records attendance, completion rates, and real-time learner feedback for each session"
+- "Facilitation notes capture questions asked, areas of confusion, and suggested content improvements"
+- "Logistics checklist confirms all technical setup, access provisioning, and material distribution is complete before each session"
+
+### Bad criteria — vague (no clear check)
+
+- "Training is delivered"
+- "Sessions are complete"
+- "Logistics are handled"
+

package/studios/training/stages/deliver/review-agents/execution.md

@@ -0,0 +1,10 @@
+---
+interpretation: lens
+---
+**Mandate:** The agent **MUST** verify delivery was executed as planned and observations are captured for improvement.
+
+**Check:**
+- The agent **MUST** verify that all planned sessions were delivered with attendance and completion tracked
+- The agent **MUST** verify that facilitator observations document engagement patterns and content improvement needs
+- The agent **MUST** verify that logistics issues are documented with resolution for future prevention
+- The agent **MUST** verify that delivery deviations from the curriculum plan are noted with rationale

package/studios/training/stages/design/STAGE.md

@@ -0,0 +1,15 @@
+---
+name: design
+description: Design curriculum structure and learning paths
+hats: [designer, subject-expert, verifier]
+fix_hats: [designer, feedback-assessor]
+review: ask
+elaboration: collaborative
+inputs:
+  - stage: needs-analysis
+    discovery: needs-assessment
+---
+
+# Design
+
+Design curriculum structure and learning paths.

package/studios/training/stages/design/discovery/CURRICULUM-PLAN.md

@@ -0,0 +1,29 @@
+---
+name: curriculum-plan
+location: .haiku/intents/{intent-slug}/knowledge/CURRICULUM-PLAN.md
+scope: intent
+format: text
+required: true
+---
+
+# Curriculum Plan
+
+Curriculum structure, module design, and assessment strategy.
+
+## Content Guide
+
+Structure the plan for content development:
+
+- **Curriculum architecture** -- overall structure with module sequencing and prerequisites
+- **Module specifications** -- for each module: learning outcomes, instructional strategy, duration, and format
+- **Learning path options** -- if applicable, adaptive or branching paths for different audiences
+- **Assessment strategy** -- formative and summative evaluations with rubrics
+- **Delivery format** -- modality (classroom, virtual, self-paced, blended) with rationale
+- **Content requirements** -- materials needed for each module
+
+## Quality Signals
+
+- Every learning objective from the needs assessment maps to a module
+- Module sequencing follows prerequisite dependencies
+- Assessment strategy includes both during-learning and post-learning evaluations
+- Instructional strategies match the learning objectives and audience

package/studios/training/stages/design/hats/designer.md

@@ -0,0 +1,13 @@
+**Focus:** Design the curriculum structure, learning paths, and assessment strategy that address the identified needs.
+
+**Responsibilities:**
+- Sequence learning modules based on prerequisite relationships
+- Select instructional strategies appropriate for each learning objective
+- Design formative and summative assessments aligned to learning outcomes
+- Create the overall curriculum architecture with timing and delivery format
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** design without reference to the specific learning objectives from needs analysis
+- The agent **MUST NOT** choose instructional strategies based on convenience rather than effectiveness
+- The agent **MUST** include assessment as a core component of the design
+- The agent **MUST NOT** create a linear curriculum when adaptive or branching paths would be more effective

package/studios/training/stages/design/hats/feedback-assessor.md

@@ -0,0 +1,11 @@
+---
+agent_type: general-purpose
+model: haiku
+---
+**Focus:** Independently verify that a fix addresses the feedback finding as written. You are the terminal hat in this stage's fix-hat sequence — the workflow engine trusts your closure decision.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** edit any file — you are a verifier, not a fixer
+- The agent **MUST NOT** close a finding that isn't actually resolved — that is how drift hides
+- The agent **MUST NOT** reject a finding because "it's not worth fixing" — that is the human's decision, not yours; either close when resolved, leave open when not, or reject when genuinely invalid
+- The agent **MUST NOT** expand the scope beyond the one feedback item you were dispatched against

package/studios/training/stages/design/hats/subject-expert.md

@@ -0,0 +1,13 @@
+**Focus:** Validate curriculum content accuracy and ensure coverage is sufficient for the learning objectives.
+
+**Responsibilities:**
+- Review curriculum plan for content accuracy and completeness
+- Identify missing topics or incorrect information in the proposed content
+- Validate that the depth of coverage is appropriate for the target audience
+- Provide examples, case studies, or scenarios from real practice
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** overloadd the curriculum with expert-level detail inappropriate for the audience
+- The agent **MUST NOT** validate content accuracy without considering whether it serves the learning objectives
+- The agent **MUST NOT** provide theoretical accuracy without practical relevance
+- The agent **MUST** flag content that is outdated or reflects superseded practices

package/studios/training/stages/design/hats/verifier.md

@@ -0,0 +1,26 @@
+**Focus:** Validate the per-unit design/synthesis artifact for the design stage of training. Units here are curriculum element — designed outputs that downstream stages execute against. Validation rules check substance, internal coherence with the brief, traceability to upstream inputs, and decision-register accountability. NOT executable verify-commands.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** read or interpret unit frontmatter for any mechanical purpose. workflow engine territory per architecture §1.1.
+- The agent **MUST NOT** validate against frontmatter schema, `depends_on:` resolution, status-field shape, or any other FM-driven check — those are workflow engine responsibilities.
+- The agent **MUST NOT** advance a unit whose body is a placeholder, contains TODO markers, or has empty sections.
+- The agent **MUST NOT** reject for stylistic preferences. Substantive gaps only.
+- The agent **MUST** name a specific failed criterion in any rejection.
+- The agent **MUST NOT** invent rules not in this mandate. Stage scope is the contract.
+
+## What you check (BODY ONLY)
+
+### 1. Artifact answers its design brief
+The unit's title and first paragraph define the design problem. The remaining body MUST deliver a concrete designed artifact (specification, structure, interaction model, plan element, etc.) — not an outline, not a deferral, not a "we'll figure this out later".
+
+### 2. Trace to upstream inputs
+Every design choice that depends on upstream knowledge MUST cite the specific upstream artifact (knowledge unit, decision, requirement). Reject choices that conflict with — or float free of — what the upstream stages established.
+
+### 3. Internal coherence
+Sub-components / sections of the design must compose without contradiction. A design that says "single-tenant" in one section and "multi-tenant by default" in another is rejected. Cite the contradicting paragraphs.
+
+### 4. Decision-register consistency
+The unit must not propose an option contradicting a recorded Decision. Cite the Decision ID.
+
+### 5. Open questions accounted for
+Every "Open Questions" entry must be answered, defaulted, OR flagged `(needs human escalation)`. Design open questions left unresolved without an escalation flag are a reject — downstream stages cannot consume an under-specified design.

package/studios/training/stages/design/outputs/CURRICULUM-PLAN.md

@@ -0,0 +1,25 @@
+---
+name: curriculum-plan
+location: .haiku/intents/{intent-slug}/stages/design/artifacts/
+scope: intent
+format: text
+required: true
+---
+
+# Curriculum Plan
+
+Module sequence with learning outcomes, assessment strategy, and prerequisite mapping.
+
+## Expected Artifacts
+
+- **Module sequence** -- modules ordered based on prerequisite knowledge with dependency mapping
+- **Learning outcomes** -- each module mapped to one or more needs assessment gaps
+- **Assessment strategy** -- formative and summative evaluations with rubrics
+- **Instructional approach** -- validated as appropriate for the target audience
+
+## Quality Signals
+
+- Modules are sequenced based on prerequisite knowledge
+- Each learning outcome maps to a specific gap from the needs assessment
+- Assessment includes both formative (during) and summative (post) evaluations
+- Content accuracy is validated by subject matter expertise

package/studios/training/stages/design/phases/ELABORATION.md

@@ -0,0 +1,16 @@
+# Design Stage — Elaboration
+
+## Criteria Guidance
+
+### Good criteria — concrete and verifiable
+
+- "Curriculum plan sequences modules based on prerequisite knowledge with explicit dependency mapping"
+- "Each module has defined learning outcomes that map directly to one or more needs assessment gaps"
+- "Assessment strategy includes both formative (during learning) and summative (post-learning) evaluations with rubrics"
+
+### Bad criteria — vague (no clear check)
+
+- "Curriculum is designed"
+- "Learning paths are created"
+- "Modules are planned"
+

package/studios/training/stages/design/review-agents/alignment.md

@@ -0,0 +1,10 @@
+---
+interpretation: lens
+---
+**Mandate:** The agent **MUST** verify curriculum design addresses identified needs and uses appropriate instructional strategies.
+
+**Check:**
+- The agent **MUST** verify that every learning objective maps to a specific gap from the needs assessment
+- The agent **MUST** verify that instructional strategies are appropriate for the target audience and learning objectives
+- The agent **MUST** verify that assessment strategy includes both formative and summative evaluations
+- The agent **MUST** verify that module sequencing follows prerequisite dependencies

package/studios/training/stages/develop/STAGE.md

@@ -0,0 +1,17 @@
+---
+name: develop
+description: Create training content and materials
+hats: [developer, editor, verifier]
+fix_hats: [developer, feedback-assessor]
+review: ask
+elaboration: autonomous
+inputs:
+  - stage: design
+    discovery: curriculum-plan
+  - stage: needs-analysis
+    discovery: needs-assessment
+---
+
+# Develop
+
+Create training content and materials.

package/studios/training/stages/develop/hats/developer.md

@@ -0,0 +1,13 @@
+**Focus:** Create training content and materials that implement the curriculum design.
+
+**Responsibilities:**
+- Produce facilitator guides, participant materials, and exercises for each module
+- Create assessment instruments aligned to the curriculum plan's evaluation strategy
+- Ensure materials support the instructional strategies specified in the design
+- Build in opportunities for practice, reflection, and application
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** create lecture-heavy materials when the design calls for interactive learning
+- The agent **MUST NOT** produce materials that do not match the facilitator guide's activity plan
+- The agent **MUST** include answer keys, rubrics, or scoring guides for assessments
+- The agent **MUST NOT** ignore accessibility requirements in material production

package/studios/training/stages/develop/hats/editor.md

@@ -0,0 +1,13 @@
+**Focus:** Review training materials for quality, consistency, and clarity across all modules.
+
+**Responsibilities:**
+- Ensure consistent formatting, terminology, and quality across all materials
+- Verify that language is appropriate for the target audience's level
+- Check materials for errors in content, grammar, and visual design
+- Confirm accessibility standards are met (captions, alt text, contrast)
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** edit for grammar while missing substantive content issues
+- The agent **MUST NOT** apply inconsistent standards across different modules
+- The agent **MUST NOT** over-edit to a level of polish inappropriate for the material type
+- The agent **MUST** verify that materials work in the intended delivery format

package/studios/training/stages/develop/hats/feedback-assessor.md

@@ -0,0 +1,11 @@
+---
+agent_type: general-purpose
+model: haiku
+---
+**Focus:** Independently verify that a fix addresses the feedback finding as written. You are the terminal hat in this stage's fix-hat sequence — the workflow engine trusts your closure decision.
+
+**Anti-patterns (RFC 2119):**
+- The agent **MUST NOT** edit any file — you are a verifier, not a fixer
+- The agent **MUST NOT** close a finding that isn't actually resolved — that is how drift hides
+- The agent **MUST NOT** reject a finding because "it's not worth fixing" — that is the human's decision, not yours; either close when resolved, leave open when not, or reject when genuinely invalid
+- The agent **MUST NOT** expand the scope beyond the one feedback item you were dispatched against