tribunal-kit 8.0.0 → 9.0.0

package/.agent/skills/playwright-ai-e2e/SKILL.md

@@ -2,8 +2,8 @@
 name: playwright-ai-e2e
 description: Modern Playwright 1.45+ E2E web testing, resilient ARIA locators, visual regression testing, network mocking, and AI-assisted flakiness detection.
 tools: Read, Grep, Glob, Edit, Write
-version: 3.0.0
-last-updated: 2026-08-05
+version: 4.0.0
+last-updated: 2026-09-07
 script: .agent/scripts/test_runner.js
 scripts-binding:
   - .agent/scripts/test_runner.js
@@ -24,6 +24,12 @@ Before writing end-to-end web tests:
 2. Auto-Waiting & Zero Sleep → Avoid `page.waitForTimeout()`; rely on Playwright built-in auto-waiting
 3. Network Interception → Mock external third-party APIs using `page.route()` for deterministic CI runs
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Modern Playwright 1.45+ E2E web testing, resilient ARIA locators, visual regression testing, network mocking, and AI-assisted flakiness detection..
+- **DO NOT activate when:** The task falls strictly outside playwright-ai-e2e domain or belongs to a different dedicated specialist.
+
 ## Resilient E2E API Route Mocking & Interaction Test
 
 ```typescript
@@ -54,7 +60,27 @@ test.describe('Dashboard Features', () => {
 });
 ```
 
-## 🛑 Verification-Before-Completion (VBC) Protocol
+---
+
+## 🏛️ 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
+```
+✅ 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?
+```
 
-- Run Playwright test suite in headless mode and verify zero flakiness.
-- Ensure all interactive elements rely on resilient locators.
+### 🛑 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/playwright-best-practices/SKILL.md

@@ -2,8 +2,8 @@
 name: playwright-best-practices
 description: Playwright End-to-End (E2E) testing mastery. Resilient selectors, auto-waiting mechanisms, parallel test execution, mocking network requests, fixture management, and cross-browser CI configurations. Use when configuring, deploying, or writing E2E web tests.
 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:
   - testing-patterns
   - webapp-testing
@@ -25,6 +25,12 @@ Before writing Playwright E2E tests or test fixtures, you MUST inspect:
 2. Complete Test Isolation Rule (Section 53) → Isolate every test in fresh browser context; ban cascading tests dependent on previous test outcomes
 3. External Network Interception (Section 86) → Intercept 3rd party external APIs (`page.route()`) to eliminate flaky network failures in CI
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Playwright End-to-End (E2E) testing mastery. Resilient selectors, auto-waiting mechanisms, parallel test execution, mocking network requests, fixture management, and cross-browser CI configurations. Use when configuring, deploying, or writing E2E web tests..
+- **DO NOT activate when:** The task falls strictly outside playwright-best-practices domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ Using `page.waitForTimeout(3000)` for synchronization -> ✅ Use `page.waitForSelector()`, `expect(locator).toBeVisible()`, or auto-waiting locators
@@ -34,8 +40,6 @@ Before writing Playwright E2E tests or test fixtures, you MUST inspect:
 
 ---
 
-# Playwright E2E — Bulletproof Testing Mastery
-
 ---
 
 ## 1. Resilience & Auto-Waiting
@@ -167,69 +171,17 @@ 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?
@@ -238,8 +190,6 @@ 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/polish/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: polish
 description: Final production quality pass for spacing, alignment, visual rhythm, dark mode consistency, and edge states. Use before shipping or merging a feature to ensure 100% UI fidelity.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - better-ui
   - harden
@@ -27,6 +27,12 @@ Before running a pre-ship quality pass, you MUST inspect:
 
 The comprehensive pre-ship inspection checklist to verify that code meets top-tier design-engineering standards.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Final production quality pass for spacing, alignment, visual rhythm, dark mode consistency, and edge states. Use before shipping or merging a feature to ensure 100% UI fidelity..
+- **DO NOT activate when:** The task falls strictly outside polish domain or belongs to a different dedicated specialist.
+
 ---
 
 ## Pre-Ship 10-Point Polish Checklist
@@ -46,28 +52,25 @@ The comprehensive pre-ship inspection checklist to verify that code meets top-ti
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Shipping without truncation**: Letting long user inputs overflow card boundaries or break layouts.
-2. **Missing Dark Mode Contrast**: Using dark text on dark surfaces in theme toggles.
-
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
-
-**Active reviewers: `frontend-reviewer` · `ui-ux-auditor` · `accessibility-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
-
 ```
-✅ Have all 10 polish items been verified across light and dark modes?
-✅ Is text truncation handled for edge-case content lengths?
-✅ Do interactive elements show visible focus indicators during keyboard tab testing?
+✅ 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/powershell-windows/SKILL.md

@@ -2,8 +2,8 @@
 name: powershell-windows
 description: PowerShell and Windows environment mastery. Object-oriented piping, strict error handling (ErrorActionPreference), PSProviders, active directory querying, credential management, and execution policies. Use when automating Azure, Windows environments, or writing .ps1 scripts.
 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:
   - server-management
   - devops-engineer
@@ -25,7 +25,11 @@ Before writing PowerShell `.ps1` scripts or Windows automation commands, you MUS
 2. Mandatory Script Strict Mode Header (Section 43) → Always declare `$ErrorActionPreference = "Stop"` and `Set-StrictMode -Version Latest` at top of automation scripts
 3. Process-Scoped Execution Policy Bypass (Section 68) → Use process-level policy override (`powershell.exe -ExecutionPolicy Bypass -File ...`); ban system-wide `Set-ExecutionPolicy Unrestricted`
 
-# PowerShell — Windows Automation Mastery
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring PowerShell and Windows environment mastery. Object-oriented piping, strict error handling (ErrorActionPreference), PSProviders, active directory querying, credential management, and execution policies. Use when automating Azure, Windows environments, or writing .ps1 scripts..
+- **DO NOT activate when:** The task falls strictly outside powershell-windows domain or belongs to a different dedicated specialist.
 
 ---
 
@@ -141,69 +145,17 @@ 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?
@@ -212,8 +164,6 @@ 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/pricing-page/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: pricing-page
 description: SaaS pricing table architecture, billing cycle toggles (Monthly/Annual), feature comparison matrices, tier highlighting, and conversion optimization.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - landing-page
   - compact-landing
@@ -27,6 +27,12 @@ Before engineering SaaS pricing pages or comparison tables, you MUST inspect:
 
 Architect high-converting SaaS pricing tables with clear billing toggles, tier differentiation, and feature comparison matrices.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring SaaS pricing table architecture, billing cycle toggles (Monthly/Annual), feature comparison matrices, tier highlighting, and conversion optimization..
+- **DO NOT activate when:** The task falls strictly outside pricing-page domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 4 Pricing Table Rules
@@ -55,28 +61,25 @@ Architect high-converting SaaS pricing tables with clear billing toggles, tier d
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Jittering Price Numeral Transitions**: Price numbers shifting layout width when toggling monthly/annual billing. Use `font-variant-numeric: tabular-nums`.
-2. **Identical Visual Treatment**: Making all 3 pricing tiers look identical, preventing users from spotting the recommended plan.
-
----
+**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 "Most Popular / Pro" tier visually distinct with primary CTA styling?
-✅ Is `tabular-nums` applied to price numerals to prevent toggle layout shifts?
-✅ Is the Annual discount savings badge clearly visible next to the billing toggle?
+✅ 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/product-aware-heuristics/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: product-aware-heuristics
 description: Defines specific visual and structural design heuristics and React/Tailwind templates tailored to different product types (SaaS, developer tools, AI interfaces, landing pages, fintech, e-commerce, and editorial).
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - ui-reasoning-engine
   - interface-design
@@ -28,6 +28,12 @@ Before selecting layout templates or component heuristics, you MUST inspect:
 
 An interface must match its domain context. A developer tool should not look like a luxury editorial website. This skill defines the heuristics, layout requirements, and drop-in layout templates for different product categories.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Defines specific visual and structural design heuristics and React/Tailwind templates tailored to different product types (SaaS, developer tools, AI interfaces, landing pages, fintech, e-commerce, and editorial)..
+- **DO NOT activate when:** The task falls strictly outside product-aware-heuristics domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 1. SaaS & Enterprise Applications
@@ -397,12 +403,25 @@ export function AIChatInput({ onSubmit }) {
 
 ---
 
-## Pre-Flight Checklist
+## 🏛️ Tribunal Verification & Guardrails
 
-- [ ] Have I verified the visual guidelines match the target Product Category?
-- [ ] Have I checked for tabular numbers alignment in Fintech interfaces?
-- [ ] Have I checked for monospace format in Developer Tools?
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## VBC Protocol (Verification-Before-Completion)
+### ❌ 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 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?
+```
 
-You MUST verify existing code signatures and variables before attempting to modify or call them. No hallucination is permitted.
+### 🛑 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/progressive-blur/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: progressive-blur
 description: Smooth progressive backdrop blurs, depth overlays, and modern glassmorphism using CSS mask-image and multi-layered backdrop filters without GPU performance drops.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - 60fps-animation
   - better-colors
@@ -27,6 +27,12 @@ Before engineering glassmorphism or backdrop blurs, you MUST inspect:
 
 Architect high-performance, progressive backdrop blurs and frosted glass interfaces that feel tactile and fluid.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Smooth progressive backdrop blurs, depth overlays, and modern glassmorphism using CSS mask-image and multi-layered backdrop filters without GPU performance drops..
+- **DO NOT activate when:** The task falls strictly outside progressive-blur domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 3 Progressive Blur Techniques
@@ -87,28 +93,25 @@ Backdrop blurs can cause frame drops if animated directly.
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Animating `backdrop-filter` radius**: Animating `blur(0px)` to `blur(20px)` on every frame, which tanks GPU performance to under 20fps.
-2. **Missing vendor prefixes**: Forgetting `-webkit-backdrop-filter` for Safari compatibility.
-
----
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
-
-**Active reviewers: `frontend-reviewer` · `performance-optimizer`**
+### ❌ 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 `-webkit-backdrop-filter` vendor prefixes present?
-✅ Is `opacity` animated instead of directly interpolating `backdrop-filter` radius?
-✅ Is dark mode contrast preserved over frosted glass surfaces?
+✅ 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/project-idioms/SKILL.md

@@ -5,8 +5,8 @@ description: >
   from the developer's own code decisions. Generated by skill_evolution.js.
   Commit this file to share your Engineering Culture across the team.
   Every agent MUST respect these idioms above generic defaults.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - clean-code
   - codebase-design
@@ -32,9 +32,13 @@ Before generating code or suggesting architectural patterns, you MUST inspect:
 - ❌ Overriding project idioms without explicit justification -> ✅ Idioms represent team decisions; flag deviations with reasoning
 - ❌ Applying idioms from one project to a different project -> ✅ Idioms are project-specific; verify they apply to the current codebase
 
----
 
-# Project Idioms — Auto-Evolved Skill
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring >.
+- **DO NOT activate when:** The task falls strictly outside project-idioms domain or belongs to a different dedicated specialist.
+
+---
 
 > **Authority Level: ABSOLUTE**
 > These idioms were extracted from this project's actual code decisions.
@@ -114,69 +118,17 @@ 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?
@@ -185,8 +137,6 @@ 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.