tribunal-kit 4.2.0 → 4.3.1

package/.agent/skills/app-builder/templates/chrome-extension/TEMPLATE.md

@@ -89,4 +89,4 @@ project-name/
 - Use type-safe messaging
 - Wrap Chrome APIs in promises
 - Minimize permissions
-- Handle offline gracefully
+- Handle offline gracefully

package/.agent/skills/app-builder/templates/cli-tool/TEMPLATE.md

@@ -85,4 +85,4 @@ npm publish
 - Support both interactive and non-interactive modes
 - Use consistent output styling
 - Validate inputs with Zod
-- Exit with proper codes (0 success, 1 error)
+- Exit with proper codes (0 success, 1 error)

package/.agent/skills/app-builder/templates/electron-desktop/TEMPLATE.md

@@ -85,4 +85,4 @@ project-name/
 - Type-safe IPC with typed handlers
 - Custom title bar for native feel
 - Handle window state (maximize, minimize)
-- Auto-updates with electron-updater
+- Auto-updates with electron-updater

package/.agent/skills/app-builder/templates/express-api/TEMPLATE.md

@@ -80,4 +80,4 @@ project-name/
 - Validate all inputs with Zod
 - Centralized error handling
 - Environment-based config
-- Use Prisma for type-safe DB access
+- Use Prisma for type-safe DB access

package/.agent/skills/app-builder/templates/flutter-app/TEMPLATE.md

@@ -87,4 +87,4 @@ project_name/
 - Riverpod for state, React Query pattern for server state
 - Freezed for immutable data classes
 - Go Router for declarative navigation
-- Material 3 theming
+- Material 3 theming

package/.agent/skills/app-builder/templates/monorepo-turborepo/TEMPLATE.md

@@ -87,4 +87,4 @@ project-name/
 - Shared configs in packages/config
 - Shared types in packages/types
 - Internal packages with `workspace:*`
-- Use Turbo remote caching for CI
+- Use Turbo remote caching for CI

package/.agent/skills/app-builder/templates/nextjs-fullstack/TEMPLATE.md

@@ -119,4 +119,4 @@ project-name/
 - **Mutations**: Use Server Actions combined with React 19's `useActionState` to handle loading and error states instead of manual useState.
 - **Type Safety**: Share Zod schemas between Server Actions (input validation) and Client Forms.
 - **Security**: Always validate input data with Zod before passing it to Prisma.
-- **Styling**: Use native CSS variables in Tailwind v4 for easier dynamic theming.
+- **Styling**: Use native CSS variables in Tailwind v4 for easier dynamic theming.

package/.agent/skills/app-builder/templates/nextjs-saas/TEMPLATE.md

@@ -119,4 +119,4 @@ project-name/
 7. Run project:
    ```bash
    npm run dev
-   ```
+   ```

package/.agent/skills/app-builder/templates/nextjs-static/TEMPLATE.md

@@ -166,4 +166,4 @@ export const metadata: Metadata = {
 - **React Server Components (RSC)**: Default all components to Server Components. Only add `'use client'` when you need state (`useState`) or event listeners (`onClick`).
 - **Image Optimization**: Use the `<Image />` component but remember `unoptimized: true` for static export or use an external image CDN (Cloudinary/Imgix).
 - **Font Optimization**: Use `next/font` (Google Fonts) to automatically host fonts and prevent layout shift.
-- **Responsive**: Mobile-first design using Tailwind prefixes like `sm:`, `md:`, `lg:`.
+- **Responsive**: Mobile-first design using Tailwind prefixes like `sm:`, `md:`, `lg:`.

package/.agent/skills/app-builder/templates/nuxt-app/TEMPLATE.md

@@ -131,4 +131,4 @@ project-name/
   ```
 - **Data Fetching**: Sử dụng `useFetch` với `server: false` cho các tác vụ client-only, hoặc dùng Server Functions để type-safety tốt hơn.
 - **State**: Dùng `defineStore` (Pinia) cho global state, `useState` của Nuxt cho state đơn giản chia sẻ giữa Server/Client.
-- **Type Safety**: Tự động tạo type cho API routes (`$fetch` typed automatically).
+- **Type Safety**: Tự động tạo type cho API routes (`$fetch` typed automatically).

package/.agent/skills/app-builder/templates/python-fastapi/TEMPLATE.md

@@ -80,4 +80,4 @@ project-name/
 - Pydantic v2 for validation
 - SQLAlchemy 2.0 async sessions
 - Alembic for migrations
-- pytest-asyncio for tests
+- pytest-asyncio for tests

package/.agent/skills/app-builder/templates/react-native-app/TEMPLATE.md

@@ -116,4 +116,4 @@ project-name/
 - **React 19**: Reduce usage of `useMemo` or `useCallback` thanks to React Compiler (if enabled).
 - **Components**: Build UI primitives (Box, Text) with NativeWind className for reusability.
 - **Assets**: Use `expo-image` instead of default `<Image />` for better caching and performance.
-- **API**: Always wrap API calls with TanStack Query, avoid direct calls in `useEffect`.
+- **API**: Always wrap API calls with TanStack Query, avoid direct calls in `useEffect`.

package/.agent/skills/appflow-wireframe/SKILL.md

@@ -97,7 +97,7 @@ Before writing React, chart exactly what the user can do on the screen and what
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -109,7 +109,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -120,7 +120,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -134,4 +134,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/architecture/SKILL.md

@@ -171,7 +171,7 @@ Framework:      Next.js API   NestJS        Microservices
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -183,7 +183,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -194,7 +194,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -208,4 +208,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/authentication-best-practices/SKILL.md

@@ -141,7 +141,7 @@ function canEditPost(user: User, post: Post): boolean {
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -153,7 +153,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -164,7 +164,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -178,4 +178,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/bash-linux/SKILL.md

@@ -131,7 +131,7 @@ done
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -143,7 +143,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -154,7 +154,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -168,4 +168,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

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

@@ -175,7 +175,7 @@ Pre-Flight:  ✅ All checks passed
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -187,7 +187,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -198,7 +198,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -212,4 +212,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/brainstorming/SKILL.md

@@ -438,7 +438,7 @@ STEP 4: Format Output
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -450,7 +450,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -461,7 +461,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -475,4 +475,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/building-native-ui/SKILL.md

@@ -154,7 +154,7 @@ export function FastList({ data }) {
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -166,7 +166,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -177,7 +177,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -191,4 +191,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/clean-code/SKILL.md

@@ -333,7 +333,7 @@ function processUser(user: User) {
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -345,7 +345,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -356,7 +356,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -370,4 +370,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/code-review-checklist/SKILL.md

@@ -146,7 +146,7 @@ Evidence:    [link to terminal output, test result, or file diff]
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -158,7 +158,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -169,7 +169,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -183,4 +183,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/config-validator/SKILL.md

@@ -117,7 +117,7 @@ When automating updates to a JSON configuration (e.g., adding a new skill to `co
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -129,7 +129,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -140,7 +140,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -154,4 +154,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.

package/.agent/skills/csharp-developer/SKILL.md

@@ -470,7 +470,7 @@ public class UsersApiTests(WebApplicationFactory<Program> factory)
 
 ---
 
-## 🤖 LLM-Specific Traps
+
 
 AI coding assistants often fall into specific bad habits when dealing with this domain. These are strictly forbidden:
 
@@ -482,7 +482,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 
 ---
 
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+
 
 **Slash command: `/review` or `/tribunal-full`**
 **Active reviewers: `logic-reviewer` · `security-auditor`**
@@ -493,7 +493,7 @@ AI coding assistants often fall into specific bad habits when dealing with this
 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:
 ```
@@ -507,4 +507,12 @@ Review these questions before confirming output:
 
 **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.
+- ✅ **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.