tribunal-kit 7.0.0 → 9.0.0

package/.agent/skills/behavioral-modes/SKILL.md

@@ -2,8 +2,8 @@
 name: behavioral-modes
 description: AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type.
 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:
   - fabel-protocol
   - agentic-patterns
@@ -25,6 +25,12 @@ Before selecting an operational mode, you MUST inspect:
 2. Mode Leakage Mitigation rules (Section 160) → Prevent writing unapproved code during DISCOVER or REVIEW modes
 3. Output format contracts → Ensure findings use labeled severity tags (CRITICAL / WARNING / SUGGESTION)
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type..
+- **DO NOT activate when:** The task falls strictly outside behavioral-modes domain or belongs to a different dedicated specialist.
+
 ---
 
 ## Overview
@@ -216,69 +222,17 @@ 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?
@@ -287,8 +241,6 @@ 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/better-colors/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: better-colors
 description: OKLCH color space for web projects. Convert hex/rgb/hsl to oklch, generate palettes, check contrast, handle gamut boundaries, and theme with Tailwind v4. Triggers on oklch, color conversion, palette generation, contrast ratio, gamut, display p3, design tokens, hue drift, chroma, dark mode colors.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - better-ui
   - baseline-ui
@@ -27,6 +27,12 @@ Before defining color palettes or CSS theme variables, you MUST inspect:
 
 Design engineering guidelines for building perceptually uniform, WCAG 2.2 AA compliant, P3-gamut color systems using the modern **OKLCH** color space.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring OKLCH color space for web projects. Convert hex/rgb/hsl to oklch, generate palettes, check contrast, handle gamut boundaries, and theme with Tailwind v4. Triggers on oklch, color conversion, palette generation, contrast ratio, gamut, display p3, design tokens, hue drift, chroma, dark mode colors..
+- **DO NOT activate when:** The task falls strictly outside better-colors domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 1. Why OKLCH Over HSL / RGB
@@ -123,23 +129,25 @@ To achieve 4.5:1 text contrast against a surface:
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Using HSL for Dark Mode Inversion**: Inverting `lightness` in HSL causes dramatic hue shifts and contrast failures.
-2. **Exceeding Display P3 Gamut**: Setting Chroma `C > 0.32` without fallback causes clipping or dull rendering on non-P3 displays.
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
----
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+### ❌ 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 color custom properties defined using OKLCH?
-✅ Does text meet the minimum 4.5:1 WCAG AA contrast ratio?
-✅ Are dark mode surfaces subtly tinted rather than pure #000000 black?
+✅ 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
-
-Inspect colors across light and dark modes in browser preview to verify contrast and visual harmony.
+**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/better-typography/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: better-typography
 description: Web typography from choosing fonts to spacing, wrapping, and accessibility. Use when picking or pairing typefaces, configuring variable fonts or OpenType features, setting up a type scale, styling text in components, truncating text, styling underlines, selection, placeholders, or carets. Triggers on typography, fonts, variable fonts, font-weight, opentype, letter-spacing, line-height, type scale, tabular numbers, text-wrap, truncation, line clamp, measure, line length.
-version: 1.0.0
-last-updated: 2026-07-22
+version: 4.0.0
+last-updated: 2026-09-07
 applies-to-model: gemini-3-6-flash, claude-3-7-sonnet
 routing:
   domain: UI Craft & Design Engineering
@@ -21,10 +21,20 @@ routing:
         optical sizing,
       ]
     weak: [fonts, text style, line height]
+scripts-binding:
+  - .agent/scripts/lint_runner.js
+  - .agent/scripts/verify_all.js
 ---
 
 # Better Typography — Web Typography & Font Engineering
 
+
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Web typography from choosing fonts to spacing, wrapping, and accessibility. Use when picking or pairing typefaces, configuring variable fonts or OpenType features, setting up a type scale, styling text in components, truncating text, styling underlines, selection, placeholders, or carets. Triggers on typography, fonts, variable fonts, font-weight, opentype, letter-spacing, line-height, type scale, tabular numbers, text-wrap, truncation, line clamp, measure, line length..
+- **DO NOT activate when:** The task falls strictly outside better-typography domain or belongs to a different dedicated specialist.
+
 Design engineering guidelines for setting up modern, readable, scalable typography with OpenType features, fluid type scales, and optimal measure.
 
 ---
@@ -103,23 +113,25 @@ body {
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Using Pixel Font Sizes**: Defining `font-size: 14px` instead of `0.875rem` breaks user browser accessibility font size preferences.
-2. **Missing `text-wrap` Support**: Forgetting `text-wrap: balance` on multi-line card titles.
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
----
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+### ❌ 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 body paragraph containers capped at max-width: 65ch?
-✅ Is text-wrap: balance applied to headings?
-✅ Are tabular numbers enabled on price grids and timer counters?
+✅ 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
-
-Inspect typography scaling and line wrapping across desktop and mobile viewports in browser preview.
+**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/better-ui/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: better-ui
 description: Design engineering principles for making interfaces feel polished. Use when building UI components, reviewing frontend code, implementing animations, hover states, shadows, borders, micro-interactions, or visual detail work.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - baseline-ui
   - better-colors
@@ -21,11 +21,44 @@ scripts-binding:
 
 Before engineering component polish or micro-interactions, you MUST inspect:
 
-1. Outer vs Inner Radius Formula (Section 32) → Enforce $\text{Radius}_{\text{outer}} = \text{Radius}_{\text{inner}} + \text{Padding}_{\text{inner}}$
-2. Transition Rules (Section 25) → Strictly prohibit `transition: all`; specify explicit property transitions
-3. Multi-Layer Shadows (Section 42) → Use ambient + direct lighting stacks and dark mode surface elevation highlights
-
-Distilled design engineering principles for transforming functional UIs into polished, state-of-the-art software interfaces, based on Jakub Krehel's _Details that make interfaces feel better_.
+1. Nested Radius Formula → Enforce `Radius(outer) = Radius(inner) + Padding(inner)` to prevent visual corner clipping
+2. Explicit Property Transitions → Strictly ban `transition: all`; specify explicit properties (`transform`, `opacity`, `box-shadow`)
+3. OKLCH Color Space & Contrast → Use perceptual OKLCH color palettes; verify text contrast against APCA / WCAG 2.2 standards
+4. Anti-Slop Directive → Ban generic purple/indigo AI gradients and default left-text/right-image hero templates
+
+## Activation Boundaries
+
+- **Activate when:** Styling frontend components, refining layout spacing, crafting hover/press states, elevating visual depth/shadows, and deslopping generic UI templates.
+- **DO NOT activate when:** Writing backend database schemas, SQL queries, or infrastructure deployment scripts.
+
+## 2026 UI Design Engineering Invariants
+
+1. **The Nested Border Radius Formula**:
+   ```css
+   /* Inner card radius = 8px, Container padding = 16px -> Outer container radius = 24px */
+   .card-container { padding: 16px; border-radius: 24px; }
+   .card-inner     { border-radius: 8px; }
+   ```
+2. **Perceptual OKLCH Color Scale**:
+   ```css
+   :root {
+     --surface-0: oklch(0.14 0.01 260);
+     --surface-1: oklch(0.18 0.01 260);
+     --primary:   oklch(0.65 0.18 145); /* High-chroma, uniform lightness */
+   }
+   ```
+3. **Subgrid for Pixel-Perfect Card Alignment**:
+   ```css
+   .card-grid { display: grid; grid-template-columns: repeat(3, 1fr); }
+   .card-grid > .card { display: grid; grid-template-rows: subgrid; grid-row: span 3; }
+   ```
+
+## Hallucination Traps (Read First)
+
+- ❌ `transition: all 0.3s ease` → ✅ Specify explicit properties `transition: transform 150ms ease, opacity 150ms ease`
+- ❌ Hardcoded single muddy black shadows → ✅ Multi-layer ambient + directional shadows
+- ❌ Mismatched concentric border radii → ✅ Outer radius MUST equal inner radius + padding
+- ❌ Generic AI violet/purple mesh gradients → ✅ Use subtle grain, solid contrast, and refined OKLCH accents
 
 ---
 
@@ -96,23 +129,25 @@ Distilled design engineering principles for transforming functional UIs into pol
 
 ---
 
-## 🤖 LLM-Specific Traps
-
-1. **`transition: all` Laziness**: Causes layout recalculations and flashes during dark/light mode toggles.
-2. **Missing Active States**: Adding hover effects but forgetting `:active` click feedback.
+## 🏛️ Tribunal Verification & Guardrails
 
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+### ❌ 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 inner vs outer radii mathematically aligned?
-✅ Did I use specific CSS transition properties instead of transition: all?
-✅ Is hover and active state feedback explicit on all interactive targets?
+✅ 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
-
-Inspect micro-interactions and nested corner geometry in browser preview before delivery.
+**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/bolder/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: bolder
 description: Increase visual impact, punch, and personality for generic or bland interfaces. Use when the user asks to make the UI pop, stand out, have more character, or feel less template-like.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - better-colors
   - better-ui
@@ -27,6 +27,12 @@ Before applying bold visual transformations, you MUST inspect:
 
 Transform generic, boring template UIs into high-impact, distinctive interfaces with strong visual presence.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Increase visual impact, punch, and personality for generic or bland interfaces. Use when the user asks to make the UI pop, stand out, have more character, or feel less template-like..
+- **DO NOT activate when:** The task falls strictly outside bolder domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 5 Tactics for Bold UI Transformation
@@ -56,28 +62,25 @@ Transform generic, boring template UIs into high-impact, distinctive interfaces
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Adding Generic Purple Gradients**: Relying on generic AI gradient presets instead of distinct brand colors.
-2. **Making Everything Loud**: Applying high contrast to all text and cards, which destroys readability. Keep body copy clean and legible.
-
----
+**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
-
 ```
-✅ Is the primary accent color distinct and used intentionally?
-✅ Does the display typography maintain legibility while being high-impact?
-✅ Is body text kept crisp and high-contrast on readable backgrounds?
+✅ 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/brainstorming/SKILL.md

@@ -2,8 +2,8 @@
 name: brainstorming
 description: Socratic methodology, ideation, and architectural exploration mastery. Generating extensive feature options, analyzing trade-offs, questioning assumptions, mind-mapping components, and delaying execution. Use when evaluating new features, defining project goals, or guiding users through ambiguous design spaces.
 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:
   - plan-writing
   - fabel-protocol
@@ -21,6 +21,12 @@ Before initiating ideation sessions or generating architectural options, you MUS
 2. The Socratic Protocol (Section 1) → Provide 3 distinct options (Conservative, Balanced, Ambitious) with tabular trade-offs
 3. Final Brainstorming Assertions schema (Section 4) → Distill options into actionable requirement items before planning
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Socratic methodology, ideation, and architectural exploration mastery. Generating extensive feature options, analyzing trade-offs, questioning assumptions, mind-mapping components, and delaying execution. Use when evaluating new features, defining project goals, or guiding users through ambiguous design spaces..
+- **DO NOT activate when:** The task falls strictly outside brainstorming domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ Jumping to implementation during brainstorming -> ✅ Brainstorming is exploration only; no code is written in this phase
@@ -474,69 +480,17 @@ 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?
@@ -545,8 +499,6 @@ 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/browser-native-ai/SKILL.md

@@ -2,8 +2,8 @@
 name: browser-native-ai
 description: Browser-native AI mastery. Zero-latency local inference, ONNX Runtime Web, WebNN API hardware acceleration, WebAssembly memory boundaries, and privacy-first AI architectures.
 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:
   - generative-ui-expert
   - webgpu-performance
@@ -27,6 +27,12 @@ Before writing browser-native AI or client-side inference code, you MUST inspect
 
 You are an expert at running AI models directly on the client's device, inside the web browser. You avoid server-side APIs for privacy, cost reduction, and zero-latency execution. Your domain covers running Small Language Models (SLMs), embeddings, and vision models via ONNX Runtime Web and WebNN.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Browser-native AI mastery. Zero-latency local inference, ONNX Runtime Web, WebNN API hardware acceleration, WebAssembly memory boundaries, and privacy-first AI architectures..
+- **DO NOT activate when:** The task falls strictly outside browser-native-ai domain or belongs to a different dedicated specialist.
+
 ## 1. Core Principles
 
 - **Privacy by Default:** Data never leaves the browser. This is critical for HIPAA compliance, banking, and private notes apps.
@@ -77,42 +83,19 @@ Before submitting code, ensure:
 2. WebGPU is requested as the primary execution provider.
 3. Model payload sizes are actively considered and documented in comments.
 
-### 🛑 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.
-
 ---
 
-## 🤖 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?
@@ -121,8 +104,6 @@ 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/build-primitive/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: build-primitive
 description: Build foundational, unstyled, accessible UI primitives from scratch with strong ARIA attributes, keyboard navigation, focus traps, and state management. Use when creating custom Headless UI components (Dialog, Combobox, Accordion, Popover, Menu).
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - react-specialist
   - baseline-ui
@@ -27,6 +27,12 @@ Before building headless UI primitives, you MUST inspect:
 
 Build rock-solid, framework-agnostic or React headless UI primitives with complete ARIA pattern compliance, keyboard control, and focus management.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Build foundational, unstyled, accessible UI primitives from scratch with strong ARIA attributes, keyboard navigation, focus traps, and state management. Use when creating custom Headless UI components (Dialog, Combobox, Accordion, Popover, Menu)..
+- **DO NOT activate when:** The task falls strictly outside build-primitive domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 4 Rules for UI Primitives
@@ -57,28 +63,25 @@ Build rock-solid, framework-agnostic or React headless UI primitives with comple
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Forgetting `type="button"`**: Omitting `type="button"` on custom trigger buttons inside forms, causing unintended form submissions.
-2. **Missing `aria-expanded`**: Toggle triggers missing `aria-expanded` attributes.
-
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
-
-**Active reviewers: `accessibility-reviewer` · `type-safety`**
+### ❌ 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 WAI-ARIA roles and states correctly attached?
-✅ Is keyboard navigation fully implemented (Arrow keys, Enter, Escape)?
-✅ Is focus trapped inside modals and restored to the trigger upon exit?
+✅ 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.