tribunal-kit 8.0.0 → 9.0.0

package/.agent/skills/edge-computing/SKILL.md

@@ -2,8 +2,8 @@
 name: edge-computing
 description: Edge computing mastery. Cloudflare Workers, Vercel Edge Functions, Durable Objects, edge-compatible data patterns, cold start elimination, caching policies (Stale-While-Revalidate), and global data locality. Use when designing globally distributed, extreme low-latency applications 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:
   - nextjs-react-expert
   - local-first-architecture
@@ -25,6 +25,12 @@ Before engineering edge computing handlers or Cloudflare Workers, you MUST inspe
 2. Stale-While-Revalidate Caching (Section 47) → Use `Cache-Control: s-maxage=60, stale-while-revalidate=86400` and asynchronous `ctx.waitUntil()` cache puts
 3. Edge Database Access (Section 81) → Never open direct raw TCP connections to DBs; use HTTP connection poolers (Prisma Accelerate, LibSQL, Supabase Pooler)
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Edge computing mastery. Cloudflare Workers, Vercel Edge Functions, Durable Objects, edge-compatible data patterns, cold start elimination, caching policies (Stale-While-Revalidate), and global data locality. Use when designing globally distributed, extreme low-latency applications architectures..
+- **DO NOT activate when:** The task falls strictly outside edge-computing domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ Importing Node.js-only APIs (fs, net, child_process) in edge functions -> ✅ Edge runtime has NO Node.js APIs; use Web APIs only
@@ -33,8 +39,6 @@ Before engineering edge computing handlers or Cloudflare Workers, you MUST inspe
 
 ---
 
-# Edge Computing — Global Latency Mastery
-
 ---
 
 ## 1. The Edge Model (V8 Isolates vs Node.js)
@@ -162,69 +166,17 @@ 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?
@@ -233,8 +185,6 @@ 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/emil-design-eng/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: emil-design-eng
 description: Encodes Emil Kowalski's philosophy on UI polish, component design, animation decisions, and the invisible details that make software feel great. Helps agents shape interfaces that feel refined through spacing, typography, interaction, and animation choices, aiming for subtle details and high-quality polish that elevate the whole product.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - apple-design
   - motion-engineering
@@ -25,6 +25,12 @@ Before designing or reviewing UI micro-animations, you MUST inspect:
 2. Easing Rule (Section 75) → Use strong `ease-out` (`cubic-bezier(0.23, 1, 0.32, 1)`) for entering UI elements; ban `ease-in`
 3. Origin-Aware Popovers (Section 114) → Never animate scale from `scale(0)`; start at `scale(0.95)` with origin bound to trigger coordinates
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Encodes Emil Kowalski's philosophy on UI polish, component design, animation decisions, and the invisible details that make software feel great. Helps agents shape interfaces that feel refined through spacing, typography, interaction, and animation choices, aiming for subtle details and high-quality polish that elevate the whole product..
+- **DO NOT activate when:** The task falls strictly outside emil-design-eng domain or belongs to a different dedicated specialist.
+
 ## Initial Response
 
 When this skill is first invoked without a specific question, respond only with:
@@ -139,34 +145,25 @@ Popovers should scale in from their trigger, not from center. Explicitly set `tr
 
 ---
 
-## 🤖 LLM-Specific Traps
-
-1. **Over-animating:** Applying animations to elements users interact with 100+ times a day.
-2. **Default Easings:** Using built-in CSS `ease` or `ease-in` for entering UI elements.
-3. **Sluggishness:** Proposing UI animations longer than 300ms.
-4. **Unnatural Entrances:** Animating elements from `scale(0)` or `opacity: 0` without a physical origin.
-5. **Ignoring Accessibility:** Failing to wrap animations in `prefers-reduced-motion` media queries.
-
-## 🏛️ Tribunal Integration (Anti-Hallucination)
+## 🏛️ Tribunal Verification & Guardrails
 
 **Slash command: `/review` or `/tribunal-full`**
-**Active reviewers: `frontend-reviewer` · `accessibility-reviewer`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-### ✅ Pre-Flight Self-Audit
-
-Review these questions before confirming output:
+### ❌ 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 justify the animation's existence based on frequency?
-✅ Are all UI animations under 300ms?
-✅ Did I use `ease-out` (custom cubic-bezier) for entering elements?
-✅ Are popovers origin-aware and avoiding `scale(0)`?
-✅ Is `prefers-reduced-motion` respected?
+✅ 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 animation work complete because the code compiles.
-- ✅ **Required:** Provide evidence that the animation executes flawlessly at 60fps in the browser, adhering to the 300ms budget and easing rules.
+- ❌ **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/error-resilience/SKILL.md

@@ -2,8 +2,8 @@
 name: error-resilience
 description: Error resilience and fault tolerance mastery. Retry strategies (exponential backoff with jitter), circuit breakers, bulkheads, graceful degradation, React error boundaries, dead letter queues, timeout patterns, fallback chains, and idempotent error recovery. Use when building fault-tolerant systems, handling flaky external services, or preventing cascading failures.
 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:
   - devops-incident-responder
   - backend-security-expert
@@ -21,19 +21,38 @@ scripts-binding:
 
 Before writing retry logic, circuit breakers, or error handling routines, you MUST inspect:
 
-1. Idempotency Key Rule for Retries (Section 123) → Require explicit idempotency keys when retrying non-safe operations (POST/PUT); ban un-keyed retries
-2. Exponential Backoff with Jitter (Section 106) → Implement randomized full jitter with capped exponential delays; ban fixed-delay retry loops
-3. Operational vs Programmer Error Split (Section 28) → Catch and recover strictly from `OperationalError` instances; allow programmer bugs (`TypeError`) to crash fast
+1. Idempotency Key Rule for Retries → Require explicit idempotency keys when retrying state-changing mutations (POST/PUT); ban un-keyed blind retries
+2. Exponential Backoff with Full Jitter → Implement randomized full jitter with capped exponential delays (`sleep = rand(0, min(cap, base * 2^attempt))`); ban fixed-delay retry loops
+3. Global Request Timeout Budgets → Always pass and respect `AbortSignal.timeout(ms)` to cancel downstream calls before connection pool starvation occurs
+4. Operational vs Programmer Error Split → Catch and recover strictly from known operational errors (network blips, rate limits); fail fast on programmer bugs (`TypeError`, null references)
 
-## Hallucination Traps (Read First)
+## Activation Boundaries
 
-- ❌ Retrying non-idempotent operations (POST payments) without idempotency keys -> ✅ Always require idempotency keys for non-safe HTTP methods before retrying
-- ❌ Using fixed delay retries instead of exponential backoff -> ✅ Exponential backoff with jitter prevents thundering herd
-- ❌ Catching all errors and swallowing them -> ✅ Only catch recoverable (operational) errors; let programmer errors crash
+- **Activate when:** Building resilient distributed systems, implementing retry loops, circuit breakers, timeout budgets, bulkhead isolation, and React error boundaries.
+- **DO NOT activate when:** Writing localized pure mathematical transformations or simple CSS styling.
 
----
+## 2026 Fault Tolerance & Resilience Invariants
 
-# Error Resilience — Fault-Tolerant Systems
+1. **`AbortSignal.timeout()` Budgeting**:
+   ```ts
+   // Native in modern Node.js and browsers
+   const res = await fetch(url, { signal: AbortSignal.timeout(3000) });
+   ```
+2. **AWS Architecture Full Jitter Formula**:
+   ```ts
+   function getJitteredBackoff(attempt: number, baseMs = 100, maxCapMs = 5000): number {
+     const exponential = Math.min(maxCapMs, baseMs * 2 ** attempt);
+     return Math.random() * exponential;
+   }
+   ```
+3. **Circuit Breaker Trip Thresholds**: Trip to `OPEN` on ≥ 5 consecutive 5xx errors or > 50% failure rate over a 10s rolling window. Fail fast immediately with a cached or fallback response while open.
+
+## Hallucination Traps (Read First)
+
+- ❌ Retrying non-idempotent operations without idempotency keys → ✅ Blind retries cause double charges and duplicated records
+- ❌ Retrying indefinitely without exponential backoff → ✅ Causes the "thundering herd" problem and DDOSes reviving downstream servers
+- ❌ Swallowing errors silently `catch (e) {}` → ✅ Logs must capture the root stack trace and error cause
+- ❌ Using `setTimeout` without clearing when promise resolves → ✅ Causes timer memory leaks
 
 ---
 
@@ -407,69 +426,17 @@ 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?
@@ -478,8 +445,6 @@ 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/expo-router-v4/SKILL.md

@@ -2,8 +2,8 @@
 name: expo-router-v4
 description: React Native 0.76+ New Architecture (Fabric/TurboModules), Expo Router v4 typed file-based navigation, native haptics, and biometrics.
 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/lint_runner.js
 scripts-binding:
   - .agent/scripts/lint_runner.js
@@ -23,6 +23,12 @@ Before building mobile navigation or components:
 2. Typed Routing → Use `expo-router` typed routes for safe navigation
 3. Safe Area & Haptics → Wrap screens in `SafeAreaView` and provide subtle `expo-haptics` feedback
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring React Native 0.76+ New Architecture (Fabric/TurboModules), Expo Router v4 typed file-based navigation, native haptics, and biometrics..
+- **DO NOT activate when:** The task falls strictly outside expo-router-v4 domain or belongs to a different dedicated specialist.
+
 ## Layout Navigation Architecture (`app/_layout.tsx`)
 
 ```tsx
@@ -75,7 +81,27 @@ export function TouchButton({ label, onPress, ...props }: TouchButtonProps) {
 }
 ```
 
-## 🛑 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?
+```
 
-- Verify zero bridge warnings during navigation.
-- Ensure safe area insets are respected across iOS notch and Android gesture bars.
+### 🛑 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/extract-design-system/SKILL.md

@@ -2,8 +2,8 @@
 name: extract-design-system
 description: Design system extraction and tokenization mastery. Identifying repeated HTML/CSS patterns, extracting CSS variables, generating design tokens (colors, spacing, typography), building reusable component schemas, and standardizing ad-hoc styles into cohesive global systems. Use when refactoring messy CSS into a unified design system.
 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:
   - create-design-md
   - tailwind-patterns
@@ -25,6 +25,12 @@ Before extracting design systems or tokenizing CSS variables, you MUST inspect:
 2. Semantic Naming Convention (Section 17) → Use functional/semantic names (`--color-primary`) rather than literal colors (`--red-500`)
 3. Accessibility Token Contract (Section 110) → Ensure extracted text and surface tokens satisfy a min 4.5:1 WCAG AA contrast ratio
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Design system extraction and tokenization mastery. Identifying repeated HTML/CSS patterns, extracting CSS variables, generating design tokens (colors, spacing, typography), building reusable component schemas, and standardizing ad-hoc styles into cohesive global systems. Use when refactoring messy CSS into a unified design system..
+- **DO NOT activate when:** The task falls strictly outside extract-design-system domain or belongs to a different dedicated specialist.
+
 ## Hallucination Traps (Read First)
 
 - ❌ Creating CSS variables for every possible value -> ✅ Only tokenize values that appear 3+ times across components
@@ -33,8 +39,6 @@ Before extracting design systems or tokenizing CSS variables, you MUST inspect:
 
 ---
 
-# Extract Design System — Tokenization Mastery
-
 ---
 
 ## 1. The Token Extraction Protocol
@@ -139,69 +143,17 @@ 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?
@@ -210,8 +162,6 @@ 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/fabel-protocol/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: fabel-protocol
 description: Distilled Fabel-5 cognitive intelligence protocol. Injects epistemic reasoning, coding discipline, design evaluation cascades, and orchestration patterns into any AI model. Load this skill to make any model think, reason, code, and design like Fabel-5. Activates for complex builds, code generation, design tasks, and multi-agent orchestration.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - agentic-patterns
   - thinking-protocol
@@ -27,6 +27,12 @@ Before executing reasoning loops, you MUST inspect:
 
 > Distilled from the Fabel-5 system prompt (191KB → 2,000 tokens). Makes any model reason like Fabel-5.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Distilled Fabel-5 cognitive intelligence protocol. Injects epistemic reasoning, coding discipline, design evaluation cascades, and orchestration patterns into any AI model. Load this skill to make any model think, reason, code, and design like Fabel-5. Activates for complex builds, code generation, design tasks, and multi-agent orchestration..
+- **DO NOT activate when:** The task falls strictly outside fabel-protocol domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 1. Epistemic Reasoning Protocol
@@ -283,13 +289,25 @@ Before finalizing any response, ask yourself:
 
 ---
 
-## Pre-Flight Checklist
+## 🏛️ Tribunal Verification & Guardrails
 
-- [ ] Have I reviewed the Fabel epistemic confidence guidelines before starting?
-- [ ] Have I checked the framework version context in the workspace?
+**Slash command: `/review` or `/tribunal-full`**
+**Active reviewers: `logic-reviewer` · `security-auditor`**
 
-## VBC Protocol
+### ❌ 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.
 
-- [ ] Verify that any APIs or files to be updated are loaded and checked for stale context prior to execution.
+### ✅ 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?
+```
 
----
+### 🛑 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/fixing-accessibility/SKILL.md

@@ -1,8 +1,8 @@
 ---
 name: fixing-accessibility
 description: Audit and fix HTML accessibility issues including ARIA labels, keyboard navigation, focus management, color contrast, and form errors. Use when adding interactive controls, forms, dialogs, or reviewing WCAG compliance.
-version: 3.0.0
-last-updated: 2026-07-30
+version: 4.0.0
+last-updated: 2026-09-07
 skills:
   - audit-and-fix
   - build-primitive
@@ -27,6 +27,12 @@ Before remediating HTML accessibility defects, you MUST inspect:
 
 Guidelines for detecting, prioritizing, and fixing accessibility defects across web interfaces.
 
+
+## Activation Boundaries
+
+- **Activate when:** Operating in tasks requiring Audit and fix HTML accessibility issues including ARIA labels, keyboard navigation, focus management, color contrast, and form errors. Use when adding interactive controls, forms, dialogs, or reviewing WCAG compliance..
+- **DO NOT activate when:** The task falls strictly outside fixing-accessibility domain or belongs to a different dedicated specialist.
+
 ---
 
 ## 1. Priority Priority Matrix (WCAG 2.2 AA)
@@ -94,23 +100,25 @@ Guidelines for detecting, prioritizing, and fixing accessibility defects across
 
 ---
 
-## 🤖 LLM-Specific Traps
+## 🏛️ Tribunal Verification & Guardrails
 
-1. **Stripping Focus Outlines**: Setting `outline: none` without providing a visible `:focus-visible` replacement.
-2. **Duplicate ARIA Labels**: Putting `aria-label` on both a outer button and inner text/icon simultaneously.
+**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 icon buttons properly labeled with aria-label?
-✅ Can all interactive elements be operated using keyboard only (Tab + Space/Enter)?
-✅ Do form fields have explicit aria-describedby associations for helper and error text?
+✅ 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
-
-Test keyboard navigation (`Tab`, `Shift+Tab`, `Escape`) in browser preview before finalizing code.
+**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.