@harness-engineering/cli 1.1.1 → 1.2.0

package/dist/agents/agents/skills/claude-code/harness-architecture-advisor/SKILL.md

@@ -1,388 +0,0 @@
-# Harness Architecture Advisor
-
-> Cognitive mode: **advisory-guide**. Ask questions, surface trade-offs, present options. Do NOT execute. The human decides; you inform the decision.
-
-## When to Use
-
-- When designing a new feature, module, or system boundary
-- When choosing between architectural approaches (REST vs GraphQL, monolith vs microservice, etc.)
-- When refactoring raises questions about target architecture
-- When `on_new_feature` triggers fire and the feature touches multiple modules
-- When a Design-category error is escalated from harness-diagnostics
-- NOT for implementation (use harness-execution after the decision is made)
-- NOT for bug fixes (use harness-debugging or harness-diagnostics)
-- NOT for code review (use harness-code-review)
-
-## Core Principle
-
-**This skill advises. It does not execute.**
-
-You will research, analyze, and present options with clear trade-offs. You will not write production code, create files, or make architectural choices. The human makes the decision. You document it.
-
-If you find yourself writing implementation code, STOP. You have left advisory mode. Return to presenting options.
-
-## Process
-
-### Phase 1: DISCOVER — Understand the Problem Space
-
-**Gate: This phase requires human answers. Do not proceed to Phase 2 until the human has responded.**
-
-Ask these 5 questions. Wait for answers before proceeding.
-
-1. **What problem are you solving?** Describe the user-facing or system-facing need. Not the technical solution — the problem.
-
-2. **What are your hard constraints?** Things that cannot change: existing database, specific language/framework, compliance requirements, team size, timeline, budget.
-
-3. **What are your soft preferences?** Things you would like but could trade away: specific patterns, technology preferences, performance targets, consistency with other systems.
-
-4. **What have you already considered?** Any approaches you have thought about, tried, or rejected. Include why you rejected them if applicable.
-
-5. **What does success look like in 6 months?** How will you know this decision was correct? What would make you regret it?
-
-Record the answers verbatim. Do not paraphrase or interpret at this stage.
-
-```
-Store answers in: .harness/architecture/<topic>/discovery.md
-```
-
----
-
-### Phase 2: ANALYZE — Research the Codebase
-
-Read the codebase to understand the current state. Do not propose solutions yet — gather facts.
-
-#### Step 1: Map Existing Patterns
-
-Search for how the codebase currently handles similar concerns:
-
-- What architectural patterns are already in use? (MVC, hexagonal, event-driven, etc.)
-- How are similar features structured?
-- What conventions exist for the relevant layer (API, data, UI, infrastructure)?
-
-#### Step 2: Identify Integration Points
-
-Find where the new feature will touch existing code:
-
-- Which modules will it interact with?
-- What are the current API boundaries?
-- Are there shared data models or services?
-
-#### Step 3: Assess Technical Debt
-
-Look for existing issues that may affect the decision:
-
-- Are there known pain points in the relevant area?
-- Is there existing tech debt that one option would worsen and another would improve?
-- Are there pending migrations or deprecations?
-
-#### Step 4: Summarize Findings
-
-```markdown
-## Codebase Analysis: <topic>
-
-### Current Patterns
-
-- <pattern 1>: used in <locations>
-- <pattern 2>: used in <locations>
-
-### Integration Points
-
-- <module A>: <how it connects>
-- <module B>: <how it connects>
-
-### Technical Debt
-
-- <issue 1>: <impact on this decision>
-- <issue 2>: <impact on this decision>
-
-### Relevant Files
-
-- <path>: <why it matters>
-```
-
-```
-Store analysis in: .harness/architecture/<topic>/analysis.md
-```
-
----
-
-### Phase 3: PROPOSE — Present Options with Trade-Offs
-
-**Gate: This phase requires human choice at the end. Do not proceed to Phase 4 until the human has selected an option.**
-
-Present 2-3 architectural options. Never present only one option — a single option is not a decision, it is a directive. Never present more than 3 — too many options cause decision paralysis.
-
-#### For Each Option
-
-```markdown
-### Option [A/B/C]: <Name>
-
-**Summary:** One paragraph describing the approach.
-
-**How it works:**
-
-1. <Step 1>
-2. <Step 2>
-3. <Step 3>
-
-**Pros:**
-
-- <Advantage 1> — <why it matters given the constraints>
-- <Advantage 2> — <why it matters given the constraints>
-
-**Cons:**
-
-- <Disadvantage 1> — <severity: low/medium/high> — <mitigation if any>
-- <Disadvantage 2> — <severity: low/medium/high> — <mitigation if any>
-
-**Effort:** <Small / Medium / Large> — <rough description of what is involved>
-
-**Risk:** <Low / Medium / High> — <what could go wrong>
-
-**Best when:** <the scenario where this option is clearly the right choice>
-```
-
-#### Comparison Matrix
-
-After presenting all options, provide a direct comparison:
-
-```markdown
-| Criterion        | Option A | Option B | Option C |
-| ---------------- | -------- | -------- | -------- |
-| Complexity       |          |          |          |
-| Performance      |          |          |          |
-| Maintainability  |          |          |          |
-| Effort to build  |          |          |          |
-| Effort to change |          |          |          |
-| Risk             |          |          |          |
-| Fits constraints |          |          |          |
-```
-
-#### Recommendation
-
-State which option you would lean toward and why, but frame it as a recommendation, not a decision:
-
-```
-Based on the constraints (especially <key constraint>), I would lean toward Option <X> because <reason>.
-However, if <condition>, Option <Y> would be stronger.
-```
-
-Present the options to the human and wait for their choice.
-
-```
-Store proposal in: .harness/architecture/<topic>/proposal.md
-```
-
----
-
-### Phase 4: DOCUMENT — Write the Architecture Decision Record
-
-After the human selects an option, write a formal ADR.
-
-#### ADR Template
-
-```markdown
-# ADR-<number>: <Title>
-
-**Date:** <date>
-**Status:** Accepted
-**Deciders:** <who was involved>
-
-## Context
-
-<What is the problem or need that prompted this decision? Include relevant
-background, constraints, and the current state of the system. A reader who
-was not part of the discussion should understand why this decision was needed.>
-
-## Decision
-
-<What is the architectural decision? State it clearly and specifically.
-Include enough detail that someone could implement it without further
-discussion.>
-
-## Alternatives Considered
-
-### <Alternative 1 name>
-
-<Brief description and why it was not chosen.>
-
-### <Alternative 2 name>
-
-<Brief description and why it was not chosen.>
-
-## Consequences
-
-### Positive
-
-- <Benefit 1>
-- <Benefit 2>
-
-### Negative
-
-- <Trade-off 1> — <mitigation plan>
-- <Trade-off 2> — <mitigation plan>
-
-### Neutral
-
-- <Side effect that is neither positive nor negative>
-
-## Action Items
-
-- [ ] <Concrete next step 1> — owner: <who> — by: <when>
-- [ ] <Concrete next step 2> — owner: <who> — by: <when>
-- [ ] <Concrete next step 3> — owner: <who> — by: <when>
-```
-
-Save the ADR:
-
-```
-.harness/architecture/<topic>/ADR-<number>.md
-```
-
-Also link from the project's ADR index if one exists.
-
-## Harness Integration
-
-- Extends the human-architect model — the skill is a thinking partner, not a decision maker
-- Respects architectural constraints defined in harness.config.json
-- Outputs structured ADR that other skills can reference
-- Reads prior ADRs from `.harness/architecture/` for consistency
-
-## Success Criteria
-
-- [ ] All 5 discovery questions are asked (or explicitly deferred by human)
-- [ ] At least 2 options are presented with concrete trade-offs
-- [ ] Human makes an explicit choice before documentation proceeds
-- [ ] ADR follows the template structure with all sections filled
-- [ ] ADR references specific files or components (not abstract generalities)
-
-## Gates
-
-- **Phase 1 to Phase 2: Requires human answers.** Do not proceed to codebase analysis until the human has answered the discovery questions. Without understanding constraints, analysis is directionless.
-- **Phase 3 to Phase 4: Requires human choice.** Do not write the ADR until the human has selected an option. The ADR documents a decision, not a recommendation.
-- **Always 2-3 options.** Never present 1 option (that is a directive, not advice). Never present more than 3 (that causes paralysis).
-- **No implementation in this skill.** If you write production code, you have broken the advisory boundary. Stop and return to presenting options.
-- **Trade-offs must be honest.** Every option has downsides. If you cannot articulate the cons of an option, you do not understand it well enough to recommend it.
-
-## Escalation
-
-- **Human cannot choose between options:** Help narrow by asking which constraint matters most. If two options are genuinely equivalent, say so — flip a coin on equivalent options rather than agonizing.
-- **Analysis reveals the problem is already solved:** If the codebase already has a pattern that handles this, say so. No need to architect what already exists.
-- **Constraints are contradictory:** If hard constraints rule out all reasonable options, escalate this back to the human. Something in the constraints must give.
-- **Decision has organization-wide impact:** If the decision affects teams or systems beyond the current codebase, flag this. The decision may need a broader audience.
-
-## Examples
-
-### Example 1: API Design for a New Resource
-
-**Phase 1 — DISCOVER:**
-
-```
-1. Problem: We need to expose order history to mobile clients and third-party integrations.
-2. Hard constraints: Must work with existing PostgreSQL database. REST API already
-   serves other resources. Team of 3 backend engineers. Ship in 6 weeks.
-3. Soft preferences: Would like to use GraphQL eventually. Want pagination.
-   Want to keep response times under 200ms.
-4. Already considered: Adding REST endpoints like the other resources. Wondered
-   about GraphQL but unsure if it is worth the investment for one resource.
-5. Success in 6 months: Mobile app and 2 integrations consuming the API without
-   complaints about performance or missing data.
-```
-
-**Phase 2 — ANALYZE:**
-
-```
-Current patterns: REST with Express, controller-service-repository layers.
-  15 existing resources follow this pattern.
-Integration points: Order model joins with Users and Products.
-  Existing /users and /products endpoints.
-Tech debt: No pagination on existing endpoints — they return all records.
-  This is fine for small datasets but orders could be large.
-```
-
-**Phase 3 — PROPOSE:**
-
-Option A: REST endpoints following existing patterns. Add pagination.
-
-- Pros: Consistent with codebase, team knows how to build it, fast to ship.
-- Cons: Mobile may over-fetch (gets all fields even when it needs few), no subscription support.
-- Effort: Small (2-3 weeks).
-
-Option B: GraphQL for orders only, REST for everything else.
-
-- Pros: Mobile gets exactly the fields it needs, natural pagination with cursors, foundation for future migration.
-- Cons: Two API paradigms to maintain, team learning curve, more infrastructure (Apollo/Yoga server).
-- Effort: Medium (4-5 weeks).
-
-Option C: REST with sparse fieldsets (JSON:API style `?fields=id,total,date`).
-
-- Pros: Stays REST, solves over-fetching for mobile, no new infrastructure.
-- Cons: Custom implementation of field filtering, not a standard the team has used before, still no subscriptions.
-- Effort: Small-Medium (3-4 weeks).
-
-Recommendation: Option A with pagination. It ships fastest, matches the existing codebase, and the team can migrate to GraphQL later as a separate initiative when there are more consumers with diverse needs.
-
-**Phase 4 — DOCUMENT:**
-ADR written after human selects Option A. Action items: implement pagination middleware (reusable for other endpoints), create /orders endpoints, update API documentation.
-
-### Example 2: Component Decomposition for a Dashboard
-
-**Phase 1 — DISCOVER:**
-
-```
-1. Problem: The dashboard page component is 1200 lines and growing. It handles
-   data fetching, state management, filtering, sorting, and rendering for
-   5 different widget types.
-2. Hard constraints: React with TypeScript. Must not break existing widget
-   behavior. Cannot change the API contract — backend team is on a different
-   release cycle.
-3. Soft preferences: Want widgets to be independently testable. Want to add
-   2 new widget types next quarter without touching the main component.
-4. Already considered: Just splitting into smaller files. Worried that without
-   a clear boundary, it will re-tangle.
-5. Success in 6 months: Adding a new widget type requires creating one new
-   file and registering it, not modifying the dashboard component.
-```
-
-**Phase 2 — ANALYZE:**
-
-```
-Current patterns: Dashboard uses a single useEffect for all data fetching.
-  State is a large object with fields for each widget type. Rendering uses
-  a switch statement on widget type.
-Integration points: 3 API endpoints supply data. Shared filter context
-  affects all widgets. URL query params drive initial state.
-Tech debt: Two widget types share copy-pasted filtering logic. The sort
-  function has special cases for each widget type.
-```
-
-**Phase 3 — PROPOSE:**
-
-Option A: Widget plugin architecture with a registry.
-
-- Each widget is a self-contained module (component + hook + types).
-- A registry maps widget type strings to widget modules.
-- Dashboard iterates the registry, renders each widget, passes shared context.
-- Pros: Adding a widget means adding one module and one registry entry. Widgets are independently testable. Clear boundaries.
-- Cons: Requires upfront refactoring of all 5 existing widgets. Shared filter logic needs a common abstraction.
-- Effort: Medium (2-3 weeks to extract, then new widgets are fast).
-
-Option B: Compound component pattern with slots.
-
-- Dashboard defines layout slots. Each widget fills a slot.
-- Shared state via React context. Each widget manages its own data fetching.
-- Pros: Simpler than a registry. Familiar React patterns. Widgets own their data.
-- Cons: Less structured than a registry — no formal contract. Could re-tangle if disciplines slip. Adding a widget still requires modifying the dashboard layout.
-- Effort: Small-Medium (1-2 weeks).
-
-Option C: Micro-frontend approach with module federation.
-
-- Each widget is a separate build artifact loaded at runtime.
-- Pros: Maximum independence. Widgets can use different libraries. Independent deployment.
-- Cons: Massive overkill for 5-7 widgets in one app. Complex build setup. Runtime overhead. Team of 3 does not need this level of isolation.
-- Effort: Large (4-6 weeks).
-
-Recommendation: Option A. The plugin registry provides the clear boundary you need to prevent re-tangling, and the upfront cost pays off immediately when you add the 2 new widget types next quarter. Option C is overengineered for your scale.
-
-**Phase 4 — DOCUMENT:**
-ADR written after human selects Option A. Action items: define widget interface contract, extract existing widgets one at a time (one PR per widget), create registry with type-safe registration, add documentation for "how to add a new widget."

package/dist/agents/agents/skills/claude-code/harness-architecture-advisor/skill.yaml

@@ -1,48 +0,0 @@
-name: harness-architecture-advisor
-version: "1.0.0"
-description: Interactive architecture advisor that surfaces trade-offs and helps humans choose
-cognitive_mode: advisory-guide
-triggers:
-  - manual
-  - on_new_feature
-platforms:
-  - claude-code
-  - gemini-cli
-tools:
-  - Read
-  - Glob
-  - Grep
-  - Bash
-cli:
-  command: harness skill run harness-architecture-advisor
-  args:
-    - name: path
-      description: Project root path
-      required: false
-    - name: topic
-      description: "Architecture topic (e.g., api-design, data-modeling, decomposition)"
-      required: false
-mcp:
-  tool: run_skill
-  input:
-    skill: harness-architecture-advisor
-    path: string
-type: flexible
-phases:
-  - name: discover
-    description: Ask questions about the problem space and constraints
-    required: true
-  - name: analyze
-    description: Research the codebase and identify relevant patterns
-    required: true
-  - name: propose
-    description: Present 2-3 architectural options with trade-offs
-    required: true
-  - name: document
-    description: Write an Architecture Decision Record for the chosen option
-    required: true
-state:
-  persistent: true
-  files:
-    - .harness/architecture/
-depends_on: []

package/dist/agents/agents/skills/claude-code/harness-brainstorming/SKILL.md

@@ -1,247 +0,0 @@
-# Harness Brainstorming
-
-> Design exploration to spec to plan. No implementation before design approval. Think first, build second.
-
-## When to Use
-
-- Starting a new feature or project that requires design decisions
-- When the problem space is ambiguous and needs exploration before planning
-- When multiple implementation approaches exist and tradeoffs must be weighed
-- When `on_new_feature` trigger fires and the scope is non-trivial
-- NOT when the implementation path is already clear (go straight to harness-planning)
-- NOT when fixing a bug with an obvious root cause (use harness-debugging or harness-tdd)
-- NOT when the task is a simple refactor with no design decisions (use harness-refactoring)
-
-## Process
-
-### Iron Law
-
-**No implementation may begin before the design is approved by the human.**
-
-If you find yourself writing production code, tests, or scaffolding before the human has signed off on the design, STOP. You are in the wrong skill. Brainstorming produces a spec document, not code.
-
----
-
-### Phase 1: EXPLORE — Gather Context
-
-1. **Read the existing codebase.** Understand the current architecture, constraints, and conventions. Check AGENTS.md, existing specs in `docs/`, and relevant source files.
-
-2. **Identify the problem boundary.** What exactly needs to be solved? What is explicitly out of scope? Write down both.
-
-3. **Check for prior art.** Has this problem been partially solved elsewhere in the codebase? Are there existing patterns that should be followed or deliberately broken?
-
-4. **Assess scope.** If the problem spans more than 3 major subsystems or would take more than 2 weeks to implement, it is too large. Decompose into sub-projects first, then brainstorm each one separately.
-
----
-
-### Phase 2: EVALUATE — Ask Questions and Narrow
-
-1. **Ask ONE question at a time.** Do not dump a list of 10 questions on the human. Ask the most important question first. Wait for the answer. Let the answer inform the next question.
-
-2. **Prefer multiple choice.** Instead of "How should we handle auth?", ask "For auth, should we: (A) use existing JWT middleware, (B) add OAuth2 via provider X, or (C) delegate to an external service?" Give 2-4 concrete options with brief tradeoff notes.
-
-3. **When the human answers, acknowledge and build on it.** Do not re-ask clarified points. Track decisions as they accumulate.
-
-4. **Apply YAGNI ruthlessly.** For every proposed capability, ask: "Do we need this for the stated goal, or is this speculative?" If speculative, cut it. If the human insists, note it as a future consideration, not a requirement.
-
-5. **Continue until you have enough clarity** to propose concrete approaches. Typically 3-7 questions are sufficient. If you need more than 10, the scope is too large — decompose.
-
-### Context Keywords
-
-During Phase 2, extract 5-10 domain keywords that capture the problem space. Include them in the spec frontmatter:
-
-```
-**Keywords:** auth, middleware, session-tokens, refresh-flow, OAuth2
-```
-
-These keywords flow into the `handoff.json` `contextKeywords` field when the spec is handed off to planning. Select keywords that would help a fresh agent — one that has never seen this conversation — understand the domain area quickly.
-
----
-
-### Phase 3: PRIORITIZE — Propose Approaches
-
-1. **Propose 2-3 concrete approaches.** Not 1 (no choice), not 5 (decision paralysis). Each approach must include:
-   - **Summary:** One sentence describing the approach
-   - **How it works:** Key technical decisions (what data structures, what APIs, what patterns)
-   - **Tradeoffs:** What you gain, what you lose, what gets harder later
-   - **Estimated complexity:** Low / Medium / High, with a brief justification
-   - **Risk:** What could go wrong, what assumptions might be wrong
-
-2. **Be honest about tradeoffs.** Do not soft-sell a preferred approach. If approach A is simpler but less extensible, say so plainly.
-
-3. **State your recommendation** and why, but defer to the human's decision.
-
-4. **Wait for the human to choose.** Do not proceed until an approach is selected.
-
----
-
-### Phase 4: VALIDATE — Write the Spec
-
-1. **Present the design section by section.** Do not dump the entire spec at once. Present each major section, get feedback, and incorporate it before moving to the next:
-   - Overview and goals
-   - Decisions made (with rationale from the brainstorming conversation)
-   - Technical design (data structures, APIs, file layout)
-   - Success criteria (observable, testable outcomes)
-   - Implementation order (high-level phases, not detailed tasks — that is harness-planning's job)
-
-2. **After all sections are reviewed, write the spec to `docs/`.** Use the project's existing spec naming convention. If none exists, use `docs/specs/YYYY-MM-DD-<feature-name>.md`.
-
-   When the project has `docs/specs/`, write proposals to `docs/changes/<feature>/proposal.md` instead. This keeps change proposals separate from established specs. Fall back to the existing behavior (`docs/specs/`) when no `docs/specs/` directory exists yet.
-
-3. **Run `harness validate`** to verify the spec file is properly placed and the project remains healthy.
-
-4. **Ask for final sign-off.** Present the complete spec file path and a one-paragraph summary. The human must explicitly approve before this skill is complete.
-
----
-
-### Scope Check
-
-At any point during brainstorming, if the design reveals that the project is larger than initially expected:
-
-1. **Identify natural decomposition boundaries.** Where can the project be split into independently deliverable pieces?
-2. **Propose sub-projects.** Each sub-project should be brainstormable and plannable on its own.
-3. **Get approval for the decomposition** before continuing to brainstorm any individual sub-project.
-
-## Party Mode
-
-When activated with `--party`, add a multi-perspective evaluation step after proposing approaches.
-
-### Perspective Selection
-
-Select 2-3 perspectives based on design topic:
-
-| Topic          | Perspectives                                 |
-| -------------- | -------------------------------------------- |
-| API / backend  | Backend Developer, API Consumer, Operations  |
-| UI / frontend  | Developer, Designer, End User                |
-| Infrastructure | Architect, SRE, Developer                    |
-| Data model     | Backend Developer, Data Consumer, Migration  |
-| Library / SDK  | Library Author, Library Consumer, Maintainer |
-| Cross-cutting  | Architect, Security, Developer               |
-| Default        | Architect, Developer, User/Consumer          |
-
-### Evaluation Process
-
-For each proposed approach, evaluate from each perspective:
-
-```
-### Approach N: [name]
-
-**[Perspective 1] perspective:**
-[Assessment]. Concern: [specific concern or "None"].
-
-**[Perspective 2] perspective:**
-[Assessment]. Concern: [specific concern or "None"].
-
-**[Perspective 3] perspective:**
-[Assessment]. Concern: [specific concern or "None"].
-
-**Synthesis:** [Consensus summary. Address raised concerns. Recommend proceed/revise.]
-```
-
-Converge on a recommendation that addresses all concerns before presenting the design.
-
-## Harness Integration
-
-- **`harness validate`** — Run after writing the spec to `docs/`. Verifies project health and that the new spec file is properly placed.
-- **`harness check-docs`** — Run to verify the spec does not conflict with existing documentation.
-- **Spec location** — Specs go to `docs/` (or `docs/specs/` if the project uses that convention). Follow existing naming patterns.
-- **Handoff to harness-planning** — Once the spec is approved, invoke harness-planning to create the implementation plan from the spec.
-
-#### Requirement Phrasing
-
-When brainstorming produces requirements or acceptance criteria that will feed into planning, prefer EARS sentence patterns for behavioral requirements. See the EARS Requirement Patterns section in harness-planning for the full template reference. Key patterns:
-
-- **Event-driven:** "When [trigger], the system shall [response]."
-- **Unwanted:** "If [condition], then the system shall not [behavior]."
-
-These patterns make requirements testable and unambiguous. Apply them when the output of brainstorming includes specific behavioral expectations.
-
-## Success Criteria
-
-- A spec document exists in `docs/` with all required sections (overview, decisions, technical design, success criteria, implementation order)
-- The human explicitly approved the design before any implementation
-- YAGNI was applied: no speculative features made it into the spec
-- 2-3 approaches were presented with honest tradeoffs before a decision was made
-- Questions were asked one at a time with multiple-choice options where possible
-- `harness validate` passes after the spec is written
-- If scope was too large, it was decomposed into sub-projects with the human's approval
-
-## Examples
-
-### Example: Designing a Notification System
-
-**EXPLORE:**
-
-```
-Read AGENTS.md — project is a TypeScript monorepo with React frontend and Express backend.
-Read existing docs/ — no prior notification specs. Found docs/specs/2026-01-15-user-auth.md as naming example.
-Checked src/services/ — no notification code exists. Found email utility in src/utils/email.ts.
-Scope assessment: single subsystem (notifications), estimated 1 week. Proceed.
-```
-
-**EVALUATE:**
-
-```
-Q1: "For notifications, should we support:
-  (A) Email only (simplest, covers 90% of use cases)
-  (B) Email + in-app (covers real-time needs)
-  (C) Email + in-app + push (full coverage, most complex)?"
-Human: "B — email and in-app."
-
-Q2: "For in-app notifications, should we:
-  (A) Poll from client every 30s (simple, slight delay)
-  (B) WebSocket push (real-time, more infrastructure)?"
-Human: "A — polling is fine for now."
-
-Q3: "Should notifications be stored permanently or expire after 30 days?"
-Human: "Expire after 30 days."
-```
-
-**PRIORITIZE:**
-
-```
-Approach 1: Event-driven with queue
-- Services emit events, notification worker consumes and dispatches
-- Tradeoff: More infrastructure (needs queue), but decoupled and scalable
-- Complexity: Medium
-- Risk: Queue adds operational overhead
-
-Approach 2: Direct service calls
-- Services call NotificationService directly
-- Tradeoff: Simpler, but couples services to notification logic
-- Complexity: Low
-- Risk: Harder to add channels later, synchronous blocking
-
-Recommendation: Approach 2. YAGNI — direct calls are sufficient for email + polling.
-We can migrate to event-driven later if we add push notifications.
-Human: "Agreed, approach 2."
-```
-
-**VALIDATE:**
-
-```
-Wrote docs/specs/2026-03-14-notification-system.md
-Sections: Overview, Decisions, Technical Design, Success Criteria, Implementation Order
-harness validate — passes
-"Spec written to docs/specs/2026-03-14-notification-system.md. Ready for sign-off?"
-Human: "Approved."
-```
-
-## Gates
-
-These are hard stops. Violating any gate means the process has broken down.
-
-- **No implementation before approval.** No production code, no test code, no scaffolding, no "let me just set up the directory structure." The spec must be approved first. If you wrote code, delete it.
-- **No skipping the question phase.** You must ask at least one clarifying question before proposing approaches. If you think you already know the answer, you are making assumptions — validate them.
-- **No single-approach proposals.** Always present at least 2 approaches with tradeoffs. A single approach is a recommendation disguised as a decision — the human has no real choice.
-- **No speculative features.** Every capability in the spec must trace to a stated requirement. "We might need this later" is not a requirement. Cut it.
-- **No section-dump specs.** Present the design section by section with feedback between each. Do not write the entire spec and ask "looks good?"
-
-## Escalation
-
-- **When the human cannot decide between approaches:** Identify the key differentiator. Ask: "The main difference is X. Given your priorities, does X matter more than Y?" If still stuck, suggest building a small spike to compare (but the spike is NOT production code).
-- **When scope keeps growing during brainstorming:** Stop brainstorming. Explicitly say: "The scope has expanded beyond the original problem. Should we (A) decompose into sub-projects, or (B) revisit the original goal to narrow it?" Do not continue accumulating scope.
-- **When the problem domain is unfamiliar:** State what you do not know. Ask the human if they have domain expertise, documentation, or a reference implementation. Do not guess at domain-specific requirements.
-- **When requirements conflict with existing architecture:** Flag the conflict explicitly: "The spec calls for X, but the current architecture assumes Y. Should we (A) change the spec to work within current architecture, or (B) plan an architecture change as a prerequisite?"
-- **When you have asked more than 10 questions without converging:** The problem is too large or too ambiguous. Stop and propose a decomposition or a scoping exercise before continuing.