tribunal-kit 8.0.0 → 9.0.0

package/.agent/skills/sounds-on-the-web/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: sounds-on-the-web
 description: Web Audio API procedural sound synthesis for tactile micro-interaction feedback (clicks, pops, success chimes) with mute toggles and accessibility awareness.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - delight
   - micro-interaction
@@ -27,6 +27,12 @@ Before implementing web audio feedback, you MUST inspect:
 
 Synthesize lightweight, zero-dependency tactile sound effects for web micro-interactions.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Web Audio API procedural sound synthesis for tactile micro-interaction feedback (clicks, pops, success chimes) with mute toggles and accessibility awareness..
+- **DO NOT activate when:** The task falls strictly outside sounds-on-the-web domain or belongs to a different dedicated specialist.
+
 ---
 
 ## Zero-Asset Web Audio API Click Synthesizer
@@ -83,28 +89,25 @@ export const soundFX = new SoundFeedback();
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Autoplay Audio Errors**: Triggering Web Audio API before user gesture interaction (violating browser autoplay policies).
-2. **Missing Mute Toggle**: Failing to provide a persistent sound toggle switch in application settings.
-
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
-
-**Active reviewers: `frontend-reviewer`**
+### ❌ Forbidden AI Tropes
+1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
+2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
+3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
 ### ✅ Pre-Flight Self-Audit
-
 ```
-✅ Is AudioContext initialized strictly inside/after user gesture handlers?
-✅ Is volume gain kept subtle ($\le 0.15$ max volume)?
-✅ Is a persistent sound mute option provided?
+✅ Did I rely ONLY on real, verified tools and methods?
+✅ Is this solution appropriately scoped to the user's constraints?
+✅ Did I handle potential failure modes and edge cases?
+✅ Have I avoided generic boilerplate that doesn't add value?
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
 - ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
 - ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.

package/.agent/skills/sql-pro/SKILL.md

@@ -2,8 +2,8 @@
 name: sql-pro
 description: Senior SQL developer across major databases (PostgreSQL, MySQL, SQL Server, Oracle). Complex query design with CTEs, window functions, PIVOT, recursive queries, JSON operations, full-text search, performance optimization with EXPLAIN ANALYZE, indexing strategies, partitioning, and schema architecture. Use when writing queries, designing schemas, optimizing performance, or debugging slow queries.
 tools: Read, Grep, Glob, Bash, Edit, Write
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - database-design
   - db-latency-auditor
@@ -21,11 +21,50 @@ scripts-binding:
 
 Before writing complex SQL queries or optimizing database performance, you MUST inspect:
 
-1. SARGability Inspection (Section 328) → Ensure WHERE clauses use SARGable predicates; avoid functions or string operations on indexed columns
-2. Keyset Pagination over OFFSET (Section 74) → Use keyset/cursor pagination (`WHERE (created_at, id) < (...)`) instead of `OFFSET` for large tables (>10K rows)
-3. EXPLAIN ANALYZE Verification (Section 358) → Run `EXPLAIN (ANALYZE, BUFFERS)` to inspect physical query cost, buffer hits, and join algorithms
-
-# SQL Pro — Advanced Query & Schema Mastery
+1. Parameterized Queries Mandatory → All queries MUST use parameter markers (`$1`, `?`, `:param`); never concatenate raw user input into SQL strings
+2. SARGability Verification → Ensure `WHERE` clauses are SARGable (no functions or expressions wrapping indexed columns)
+3. Keyset Cursor Pagination → Use `WHERE (created_at, id) < ($1, $2) ORDER BY created_at DESC, id DESC LIMIT 50`; ban `OFFSET` on datasets > 1,000 rows
+4. EXPLAIN (ANALYZE, BUFFERS) → Check actual execution plan, shared buffer hits, and sequential scan flags
+
+## Activation Boundaries
+
+- **Activate when:** Designing complex SQL queries, CTEs, window functions, query performance tuning, index optimization, and debugging slow transactions.
+- **DO NOT activate when:** Writing NoSQL document queries (Mongo/Redis) or handling client-side state.
+
+## 2026 SQL Performance & Query Invariants
+
+1. **SARGable Date/Time Predicates**:
+   ```sql
+   -- ❌ BAD: Function on column kills index scan (Seq Scan)
+   WHERE DATE(created_at) = '2026-01-01'
+   -- ✅ GOOD: Direct range comparison enables Index Scan
+   WHERE created_at >= '2026-01-01 00:00:00+00' AND created_at < '2026-01-02 00:00:00+00'
+   ```
+2. **Keyset Cursor Pagination (O(1) vs O(N) OFFSET)**:
+   ```sql
+   -- ✅ Instant lookup regardless of page depth
+   SELECT id, title, created_at
+   FROM articles
+   WHERE (created_at, id) < ($last_created_at, $last_id)
+   ORDER BY created_at DESC, id DESC
+   LIMIT 20;
+   ```
+3. **CTE Optimization (`AS NOT MATERIALIZED`)**: In PostgreSQL 12+, CTEs inline automatically unless marked `MATERIALIZED`. Explicitly control materialization when CTE contains heavy calculations.
+4. **Batch Upserts with `ON CONFLICT`**:
+   ```sql
+   INSERT INTO metrics (device_id, recorded_at, val)
+   VALUES ($1, $2, $3)
+   ON CONFLICT (device_id, recorded_at)
+   DO UPDATE SET val = EXCLUDED.val;
+   ```
+
+## Hallucination Traps (Read First)
+
+- ❌ String-concatenating user input into SQL → ✅ Always use parameterized queries (`$1`, `?`)
+- ❌ Using `OFFSET 50000` for pagination → ✅ Use keyset cursor pagination (`WHERE id > $last_id`)
+- ❌ Wrapping indexed columns in functions in WHERE clauses → ✅ Compare bare column against computed value
+- ❌ `SELECT *` in production queries → ✅ Select explicit columns to utilize covering indexes
+- ❌ `COUNT(*)` on millions of rows for simple existence check → ✅ Use `SELECT EXISTS(SELECT 1 FROM ...)`
 
 ---
 
@@ -608,69 +647,17 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-**Slash command: `/review` or `/tribunal-full`**
-**Active reviewers: `logic-reviewer` · `security-auditor`**
-
-### ❌ Forbidden AI Tropes
-
-1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
-2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
-3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
-
-Review these questions before confirming output:
-
-```
-✅ Did I rely ONLY on real, verified tools and methods?
-✅ Is this solution appropriately scoped to the user's constraints?
-✅ Did I handle potential failure modes and edge cases?
-✅ Have I avoided generic boilerplate that doesn't add value?
-```
-
-### 🛑 Verification-Before-Completion (VBC) Protocol
-
-**CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
-- ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
-- ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.
-
-## Pre-Flight Checklist
-
-- [ ] Have I reviewed the user's specific constraints and requests?
-- [ ] Have I checked the environment for relevant existing implementations?
-
-## VBC Protocol (Verification-Before-Completion)
-
-You MUST verify existing code signatures and variables before attempting to modify or call them. No hallucination is permitted.
-
----
-
-## 🤖 LLM-Specific Traps
-
-AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
-
-1. **Over-engineering:** Proposing complex abstractions or distributed systems when a simpler approach suffices.
-2. **Hallucinated Libraries/Methods:** Using non-existent methods or packages. Always `// VERIFY` or check `package.json` / `requirements.txt`.
-3. **Skipping Edge Cases:** Writing the "happy path" and ignoring error handling, timeouts, or data validation.
-4. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
-5. **Silent Degradation:** Catching and suppressing errors without logging or re-raising.
-
----
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+## 🏛️ Tribunal Verification & Guardrails
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
 
 ### ❌ Forbidden AI Tropes
-
 1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
 2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
 3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
 ### ✅ Pre-Flight Self-Audit
-
-Review these questions before confirming output:
-
 ```
 ✅ Did I rely ONLY on real, verified tools and methods?
 ✅ Is this solution appropriately scoped to the user's constraints?
@@ -679,8 +666,6 @@ Review these questions before confirming output:
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
 - ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
 - ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.

package/.agent/skills/supabase-postgres-best-practices/SKILL.md

@@ -2,8 +2,8 @@
 name: supabase-postgres-best-practices
 description: Database architect expert in Supabase and PostgreSQL. Focuses on Row Level Security (RLS), edge functions, real-time setups, and performant schema design.
 tools: Read, Grep, Glob, Bash, Edit, Write
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - database-design
   - sql-pro
@@ -25,6 +25,12 @@ Before engineering Supabase schemas or writing PostgreSQL migrations, you MUST i
 2. Direct UID Match RLS Rule (Section 30) → Use direct equality (`auth.uid() = user_id`) in RLS policies; avoid slow `IN` subqueries
 3. SQL Migration File Requirement (Section 38) → Write explicit SQL migration files (`supabase/migrations/`) rather than manual GUI operations
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Database architect expert in Supabase and PostgreSQL. Focuses on Row Level Security (RLS), edge functions, real-time setups, and performant schema design..
+- **DO NOT activate when:** The task falls strictly outside supabase-postgres-best-practices domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ Using Supabase without enabling Row Level Security (RLS) -> ✅ ALL tables MUST have RLS enabled; without it, data is publicly accessible
@@ -34,8 +40,6 @@ Before engineering Supabase schemas or writing PostgreSQL migrations, you MUST i
 
 ---
 
-# Supabase & Postgres Best Practices
-
 You are a Supabase Data Architect. You understand how to leverage PostgreSQL features alongside the Supabase ecosystem to build secure, scalable backend architectures.
 
 ## Core Directives
@@ -76,69 +80,17 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-**Slash command: `/review` or `/tribunal-full`**
-**Active reviewers: `logic-reviewer` · `security-auditor`**
-
-### ❌ Forbidden AI Tropes
-
-1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
-2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
-3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
-
-Review these questions before confirming output:
-
-```
-✅ Did I rely ONLY on real, verified tools and methods?
-✅ Is this solution appropriately scoped to the user's constraints?
-✅ Did I handle potential failure modes and edge cases?
-✅ Have I avoided generic boilerplate that doesn't add value?
-```
-
-### 🛑 Verification-Before-Completion (VBC) Protocol
-
-**CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
-- ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
-- ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.
-
-## Pre-Flight Checklist
-
-- [ ] Have I reviewed the user's specific constraints and requests?
-- [ ] Have I checked the environment for relevant existing implementations?
-
-## VBC Protocol (Verification-Before-Completion)
-
-You MUST verify existing code signatures and variables before attempting to modify or call them. No hallucination is permitted.
-
----
-
-## 🤖 LLM-Specific Traps
-
-AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
-
-1. **Over-engineering:** Proposing complex abstractions or distributed systems when a simpler approach suffices.
-2. **Hallucinated Libraries/Methods:** Using non-existent methods or packages. Always `// VERIFY` or check `package.json` / `requirements.txt`.
-3. **Skipping Edge Cases:** Writing the "happy path" and ignoring error handling, timeouts, or data validation.
-4. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
-5. **Silent Degradation:** Catching and suppressing errors without logging or re-raising.
-
----
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+## 🏛️ Tribunal Verification & Guardrails
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
 
 ### ❌ Forbidden AI Tropes
-
 1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
 2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
 3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
 ### ✅ Pre-Flight Self-Audit
-
-Review these questions before confirming output:
-
 ```
 ✅ Did I rely ONLY on real, verified tools and methods?
 ✅ Is this solution appropriately scoped to the user's constraints?
@@ -147,8 +99,6 @@ Review these questions before confirming output:
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
 - ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
 - ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.

package/.agent/skills/svg-animation/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: svg-animation
 description: SVG stroke draw-on effects, path morphing, animated icons, motion paths, and interactive vector graphics.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - morphing-icons
   - 60fps-animation
@@ -27,6 +27,12 @@ Before engineering SVG path or stroke animations, you MUST inspect:
 
 Craft crisp, resolution-independent vector animations using CSS, Framer Motion, or GSAP.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring SVG stroke draw-on effects, path morphing, animated icons, motion paths, and interactive vector graphics..
+- **DO NOT activate when:** The task falls strictly outside svg-animation domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 3 SVG Motion Techniques
@@ -83,27 +89,25 @@ gsap.to('#start-shape', {
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Incorrect `stroke-dasharray` Length**: Guessing path lengths without measuring `path.getTotalLength()`, resulting in clipped or incomplete draw-on animations.
-2. **Missing `fill="none"`**: Forgetting `fill="none"` on draw-on stroke paths, causing black fill shapes during path animation.
-
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
-
-**Active reviewers: `frontend-reviewer` · `motion-reviewer`**
+### ❌ Forbidden AI Tropes
+1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
+2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
+3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
 ### ✅ Pre-Flight Self-Audit
-
 ```
-✅ Is `stroke-dasharray` calibrated to match the true SVG path length?
-✅ Is `vector-effect="non-scaling-stroke"` applied if SVG elements scale responsively?
+✅ Did I rely ONLY on real, verified tools and methods?
+✅ Is this solution appropriately scoped to the user's constraints?
+✅ Did I handle potential failure modes and edge cases?
+✅ Have I avoided generic boilerplate that doesn't add value?
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
 - ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
 - ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.

package/.agent/skills/swiftui-expert/SKILL.md

@@ -2,8 +2,8 @@
 name: swiftui-expert
 description: SwiftUI development mastery. View architecture, state management (@State, @Binding, @Environment, @Observable), performance optimization (identifiable loops, implicit vs explicit animations), architectural patterns (MVVM vs TCA), and iOS-native UX paradigms. Use when writing native Apple platforms code.
 tools: Read, Grep, Glob, Bash, Edit, Write
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - mobile-design
   - building-native-ui
@@ -25,6 +25,12 @@ Before writing or refactoring SwiftUI components, you MUST inspect:
 2. Identifiable Protocol in ForEach (Section 125) → Conform model types to `Identifiable` for `ForEach`; ban integer index iteration (`0..<items.count`)
 3. NavigationStack Standard (Section 18) → Use `NavigationStack` or `NavigationSplitView`; ban deprecated legacy `NavigationView`
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring SwiftUI development mastery. View architecture, state management (@State, @Binding, @Environment, @Observable), performance optimization (identifiable loops, implicit vs explicit animations), architectural patterns (MVVM vs TCA), and iOS-native UX paradigms. Use when writing native Apple platforms code..
+- **DO NOT activate when:** The task falls strictly outside swiftui-expert domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ Using `@State` for shared data between views -> ✅ `@State` is local to a view; use `@Binding`, `@Environment`, or `@Observable` for sharing
@@ -34,8 +40,6 @@ Before writing or refactoring SwiftUI components, you MUST inspect:
 
 ---
 
-# SwiftUI Expert — Native Apple Platforms Mastery
-
 ---
 
 ## 1. Modern State Management (iOS 17+ / Swift 5.9+)
@@ -183,69 +187,17 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-**Slash command: `/review` or `/tribunal-full`**
-**Active reviewers: `logic-reviewer` · `security-auditor`**
-
-### ❌ Forbidden AI Tropes
-
-1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
-2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
-3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
-
-Review these questions before confirming output:
-
-```
-✅ Did I rely ONLY on real, verified tools and methods?
-✅ Is this solution appropriately scoped to the user's constraints?
-✅ Did I handle potential failure modes and edge cases?
-✅ Have I avoided generic boilerplate that doesn't add value?
-```
-
-### 🛑 Verification-Before-Completion (VBC) Protocol
-
-**CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
-- ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
-- ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.
-
-## Pre-Flight Checklist
-
-- [ ] Have I reviewed the user's specific constraints and requests?
-- [ ] Have I checked the environment for relevant existing implementations?
-
-## VBC Protocol (Verification-Before-Completion)
-
-You MUST verify existing code signatures and variables before attempting to modify or call them. No hallucination is permitted.
-
----
-
-## 🤖 LLM-Specific Traps
-
-AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
-
-1. **Over-engineering:** Proposing complex abstractions or distributed systems when a simpler approach suffices.
-2. **Hallucinated Libraries/Methods:** Using non-existent methods or packages. Always `// VERIFY` or check `package.json` / `requirements.txt`.
-3. **Skipping Edge Cases:** Writing the "happy path" and ignoring error handling, timeouts, or data validation.
-4. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
-5. **Silent Degradation:** Catching and suppressing errors without logging or re-raising.
-
----
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+## 🏛️ Tribunal Verification & Guardrails
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
 
 ### ❌ Forbidden AI Tropes
-
 1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
 2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
 3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
 ### ✅ Pre-Flight Self-Audit
-
-Review these questions before confirming output:
-
 ```
 ✅ Did I rely ONLY on real, verified tools and methods?
 ✅ Is this solution appropriately scoped to the user's constraints?
@@ -254,8 +206,6 @@ Review these questions before confirming output:
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
 - ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
 - ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.

package/.agent/skills/swiss-design/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: swiss-design
 description: International Typographic Style (Swiss Design) principles for web interfaces. Strict grid discipline, asymmetric layouts, bold typographic contrast, flush-left un-justified text, and mathematical negative space.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - typeset
   - baseline-ui
@@ -27,6 +27,12 @@ Before applying Swiss Design aesthetics, you MUST inspect:
 
 Architect pristine, print-inspired interfaces grounded in grid discipline, objective clarity, and asymmetric typographic hierarchy.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring International Typographic Style (Swiss Design) principles for web interfaces. Strict grid discipline, asymmetric layouts, bold typographic contrast, flush-left un-justified text, and mathematical negative space..
+- **DO NOT activate when:** The task falls strictly outside swiss-design domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 4 Swiss Design Pillars
@@ -51,28 +57,25 @@ Architect pristine, print-inspired interfaces grounded in grid discipline, objec
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Justified Text Alignment**: Setting `text-align: justify` which creates uneven, ugly gaps ("rivers") between words. Use `text-align: left`.
-2. **Centered Text Overuse**: Centering body text paragraphs. Keep content strictly left-aligned.
-
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
-
-**Active reviewers: `frontend-reviewer` · `ui-ux-auditor`**
+### ❌ Forbidden AI Tropes
+1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
+2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
+3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
 ### ✅ Pre-Flight Self-Audit
-
 ```
-✅ Are all text paragraphs aligned flush-left (`text-align: left`)?
-✅ Is layout structured on an explicit grid axis with generous negative space?
-✅ Are tabular numbers (`tabular-nums`) used for numbered section markers?
+✅ Did I rely ONLY on real, verified tools and methods?
+✅ Is this solution appropriately scoped to the user's constraints?
+✅ Did I handle potential failure modes and edge cases?
+✅ Have I avoided generic boilerplate that doesn't add value?
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
 - ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
 - ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.

package/.agent/skills/system-design-pro/SKILL.md

@@ -2,8 +2,8 @@
 name: system-design-pro
 description: Industry-level system design mastery for interviews and production. The 6-step design framework, scale estimation (DAU→QPS→storage→bandwidth), core building blocks (load balancers L4/L7, CDN, caches, queues), database selection matrix, CAP Theorem applied, and reference designs for URL shortener, rate limiter, Twitter feed, distributed cache, and notification system. Use when designing systems for scale, conducting architecture reviews, or preparing system design discussions.
 tools: Read, Grep, Glob, Bash, Edit, Write
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - architecture
   - database-design
@@ -25,6 +25,12 @@ Before designing system architecture or writing scale proposals, you MUST inspec
 2. Quantified Scale Estimation (Section 65) → Calculate read/write QPS, storage growth, and bandwidth before selecting database or caching tiers
 3. Monolith First Rule (Section 17) → Start with modular monolith for $<10\text{K}$ RPM; ban introducing microservices or sharding prematurely
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Industry-level system design mastery for interviews and production. The 6-step design framework, scale estimation (DAU→QPS→storage→bandwidth), core building blocks (load balancers L4/L7, CDN, caches, queues), database selection matrix, CAP Theorem applied, and reference designs for URL shortener, rate limiter, Twitter feed, distributed cache, and notification system. Use when designing systems for scale, conducting architecture reviews, or preparing system design discussions..
+- **DO NOT activate when:** The task falls strictly outside system-design-pro domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ "Just use microservices" for <10K RPM → ✅ Microservices add operational complexity that kills small teams. Start monolith.
@@ -35,8 +41,6 @@ Before designing system architecture or writing scale proposals, you MUST inspec
 
 ---
 
-# System Design Pro — Industry-Level Mastery
-
 ## 1. The 6-Step Design Framework
 
 Use this structure for every system design discussion:
@@ -325,34 +329,25 @@ Failure handling:
 
 ---
 
-## 🤖 LLM-Specific Traps
-
-1. **Premature sharding**: Suggesting database sharding before the user has even mentioned scale issues. Sharding is complex; exhaust simpler options first.
-2. **NoSQL for everything**: DynamoDB and MongoDB are not universally better. They sacrifice joins and ACID. Present the tradeoffs honestly.
-3. **Ignoring the 80/20 of read QPS**: Most web apps are 80-95% reads. Design the read path first (caching, CDN, read replicas) before optimizing writes.
-4. **Forgetting the Coordinator Problem**: Distributed systems need coordination (who is the leader?). Mention ZooKeeper/etcd or leaderless designs when relevant.
-5. **Over-specifying CAP**: Real systems pick AP vs CP at the feature level, not the system level. A shopping cart (AP) and payment processing (CP) can coexist.
-
----
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+## 🏛️ Tribunal Verification & Guardrails
 
 **Slash command: `/review` or `/tribunal-full`**
-**Active reviewers: `logic-reviewer` · `system-architect` · `database-architect`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-### ✅ Pre-Flight Self-Audit
+### ❌ Forbidden AI Tropes
+1. **Blind Assumptions:** Never make an assumption without documenting it clearly with `// VERIFY: [reason]`.
+2. **Silent Degradation:** Catching and suppressing errors without logging or handling.
+3. **Context Amnesia:** Forgetting the user's constraints and offering generic advice instead of tailored solutions.
 
+### ✅ Pre-Flight Self-Audit
 ```
-✅ Did I quantify the scale before recommending a solution?
-✅ Did I present at least 2 database options with tradeoffs?
-✅ Did I define the read:write ratio before choosing a caching strategy?
-✅ Did I verify the system doesn't need microservices before recommending them?
-✅ Did I define a failure mode and mitigation for the primary bottleneck?
+✅ Did I rely ONLY on real, verified tools and methods?
+✅ Is this solution appropriately scoped to the user's constraints?
+✅ Did I handle potential failure modes and edge cases?
+✅ Have I avoided generic boilerplate that doesn't add value?
 ```
 
 ### 🛑 Verification-Before-Completion (VBC) Protocol
-
 **CRITICAL:** You must follow a strict "evidence-based closeout" state machine.
-
-- ❌ **Forbidden**: Recommending an architecture without first establishing scale numbers.
-- ✅ **Required**: Every system design must include: QPS estimates, data model, caching strategy, and at least one identified bottleneck with a scaling plan.
+- ❌ **Forbidden:** Declaring a task complete because the output "looks correct."
+- ✅ **Required:** You are explicitly forbidden from finalizing any task without providing **concrete evidence** (terminal output, passing tests, compile success, or equivalent proof) that your output works as intended.