super-opencode 1.1.0 → 1.1.2

package/.opencode/skills/self-check/SKILL.md

@@ -1,95 +1,95 @@
----
-name: self-check
-description: Post-implementation validation to enforce the Definition of Done (DoD) and prevent hallucinations.
----
-
-# Self-Check Skill
-
-## Purpose
-To validate work **after** implementation but **before** handoff.
-This is the "Check" in PDCA. It prevents the "It works on my machine" (or "It works in my context window") syndrome.
-
-**ROI Metric**: Catching a syntax error here costs 50 tokens. Catching it in a PR review costs 5,000 tokens + human time.
-
-## When to Use
--   **Trigger**: Immediately after `execution` mode finishes writing code.
--   **Trigger**: Before the `backend` agent hands off to `frontend`.
--   **Trigger**: Before any `git commit`.
-
-## The "Proof of Work" Protocol
-
-### 1. The Build Check (Syntax)
-**Rule**: *Code that doesn't build doesn't exist.*
--   Did you run the build/transpile command? (`npm run build`, `go build`, etc.)
--   **Evidence**: Paste the exit code and the last 3 lines of output.
-
-### 2. The Logic Check (Behavior)
-**Rule**: *Claims require proof.*
--   **Forbidden Phrase**: "I ran the tests and they passed." (This is often a hallucination).
--   **Required Action**: Run the test command via `run_command` and capture the actual output.
--   **Evidence**: Paste the `Total: X Passed, 0 Failed` line.
-
-### 3. The Contract Check (Requirements)
-**Rule**: *Did you solve the user's problem?*
--   Map every constraint from the original prompt to a specific line of code or test case.
-
-### 4. The Hygiene Check (Cleanup)
-**Rule**: *Leave no trace.*
--   Scan for: `console.log("here")`, `debugger`, commented out code blocks, or `TODO` comments added by you.
--   **Action**: Remove them.
-
-## Execution Template
-
-*Copy this into the context window to validate your work.*
-
-```markdown
-## ✅ Self-Check Validation
-
-### 1. Build Verification
-- **Command**: `npm run build`
-- **Output**:
-  ```text
-  > build
-  > tsc
-  ✨ Done in 2.45s.
-  ```
-- **Status**: [ PASS / FAIL ]
-
-### 2. Test Verification
-- **Command**: `npm test -- user.service.test.ts`
-- **Output**:
-  ```text
-  PASS src/services/user.service.test.ts
-  Tests: 3 passed, 3 total
-  ```
-- **Status**: [ PASS / FAIL ] -> *If FAIL, trigger `reflexion` skill.*
-
-### 3. Requirements Mapping
-| Requirement | Implemented? | Evidence (File/Line) |
-| :--- | :--- | :--- |
-| "Users must have a role" | ✅ | `UserSchema.ts:12` (Zod enum) |
-| "Passwords must be hashed" | ✅ | `AuthService.ts:45` (bcrypt call) |
-
-### 4. Code Hygiene
-- [ ] Removed debug `console.log`s?
-- [ ] Removed temporary comments?
-- [ ] Ran Linter? (`npm run lint`)
-
-### 🏁 Final Verdict: [ READY TO SHIP / NEEDS FIXING ]
-
-## Red Flags (Hallucination Detection)
-
-If you see these patterns in your own reasoning, **STOP**:
-
-| Red Flag | Why It's Bad | Correction |
-| :--- | :--- | :--- |
-| *"The code looks correct."* | Visual inspection is insufficient. | Run the code. |
-| *"I updated the tests."* | Did you *run* them? | Run the tests. |
-| *"This should fix the error."* | Hope is not a strategy. | Verify the fix. |
-| *"I'll assume the DB is running."* | Assumption is the mother of all failures. | Check connection. |
-
-## Integration with Agents
-
--   **`execution` mode**: Must run `self-check` before declaring `Status: COMPLETE`.
--   **`pm-agent`**: Rejects any task update that does not include a `self-check` artifact.
--   **`reflexion` skill**: If `self-check` results in **FAIL**, the agent must immediately enter `reflexion` mode to analyze the root cause before retrying.
+---
+name: self-check
+description: Post-implementation validation to enforce the Definition of Done (DoD) and prevent hallucinations.
+---
+
+# Self-Check Skill
+
+## Purpose
+To validate work **after** implementation but **before** handoff.
+This is the "Check" in PDCA. It prevents the "It works on my machine" (or "It works in my context window") syndrome.
+
+**ROI Metric**: Catching a syntax error here costs 50 tokens. Catching it in a PR review costs 5,000 tokens + human time.
+
+## When to Use
+-   **Trigger**: Immediately after `execution` mode finishes writing code.
+-   **Trigger**: Before the `backend` agent hands off to `frontend`.
+-   **Trigger**: Before any `git commit`.
+
+## The "Proof of Work" Protocol
+
+### 1. The Build Check (Syntax)
+**Rule**: *Code that doesn't build doesn't exist.*
+-   Did you run the build/transpile command? (`npm run build`, `go build`, etc.)
+-   **Evidence**: Paste the exit code and the last 3 lines of output.
+
+### 2. The Logic Check (Behavior)
+**Rule**: *Claims require proof.*
+-   **Forbidden Phrase**: "I ran the tests and they passed." (This is often a hallucination).
+-   **Required Action**: Run the test command via `run_command` and capture the actual output.
+-   **Evidence**: Paste the `Total: X Passed, 0 Failed` line.
+
+### 3. The Contract Check (Requirements)
+**Rule**: *Did you solve the user's problem?*
+-   Map every constraint from the original prompt to a specific line of code or test case.
+
+### 4. The Hygiene Check (Cleanup)
+**Rule**: *Leave no trace.*
+-   Scan for: `console.log("here")`, `debugger`, commented out code blocks, or `TODO` comments added by you.
+-   **Action**: Remove them.
+
+## Execution Template
+
+*Copy this into the context window to validate your work.*
+
+```markdown
+## ✅ Self-Check Validation
+
+### 1. Build Verification
+- **Command**: `npm run build`
+- **Output**:
+  ```text
+  > build
+  > tsc
+  ✨ Done in 2.45s.
+  ```
+- **Status**: [ PASS / FAIL ]
+
+### 2. Test Verification
+- **Command**: `npm test -- user.service.test.ts`
+- **Output**:
+  ```text
+  PASS src/services/user.service.test.ts
+  Tests: 3 passed, 3 total
+  ```
+- **Status**: [ PASS / FAIL ] -> *If FAIL, trigger `reflexion` skill.*
+
+### 3. Requirements Mapping
+| Requirement | Implemented? | Evidence (File/Line) |
+| :--- | :--- | :--- |
+| "Users must have a role" | ✅ | `UserSchema.ts:12` (Zod enum) |
+| "Passwords must be hashed" | ✅ | `AuthService.ts:45` (bcrypt call) |
+
+### 4. Code Hygiene
+- [ ] Removed debug `console.log`s?
+- [ ] Removed temporary comments?
+- [ ] Ran Linter? (`npm run lint`)
+
+### 🏁 Final Verdict: [ READY TO SHIP / NEEDS FIXING ]
+
+## Red Flags (Hallucination Detection)
+
+If you see these patterns in your own reasoning, **STOP**:
+
+| Red Flag | Why It's Bad | Correction |
+| :--- | :--- | :--- |
+| *"The code looks correct."* | Visual inspection is insufficient. | Run the code. |
+| *"I updated the tests."* | Did you *run* them? | Run the tests. |
+| *"This should fix the error."* | Hope is not a strategy. | Verify the fix. |
+| *"I'll assume the DB is running."* | Assumption is the mother of all failures. | Check connection. |
+
+## Integration with Agents
+
+-   **`execution` mode**: Must run `self-check` before declaring `Status: COMPLETE`.
+-   **`pm-agent`**: Rejects any task update that does not include a `self-check` artifact.
+-   **`reflexion` skill**: If `self-check` results in **FAIL**, the agent must immediately enter `reflexion` mode to analyze the root cause before retrying.

package/.opencode/skills/simplification/SKILL.md

@@ -1,92 +1,92 @@
----
-name: simplification
-description: Complexity reduction and architectural streamlining for Frontend, Backend, and Quality agents.
----
-
-# Simplification Skill (Occam's Razor)
-
-## Purpose
-"Simplicity is the ultimate sophistication." This skill provides a rigorous protocol for identifying and eliminating over-engineering, code bloat, and technical debt across the entire stack.
-
-## Overview
-The Simplification Skill is a cross-functional protocol used to flatten logic, reduce state, and ensure every line of code justifies its existence. It leverages specialized agents and MCP tools to validate architectural simplicity and performance.
-
-## Prerequisites
-- **Agents**: [Backend](file:///.opencode/agents/backend.md), [Frontend](file:///.opencode/agents/frontend.md), [Quality](file:///.opencode/agents/quality.md).
-- **MCP Tools**: 
-    - `context7`: To research lightweight alternatives and library standards.
-    - `sequential-thinking`: To audit complex logic flows for potential flattening.
-    - `sqlite`: To verify if database schemas or queries can be compressed.
-    - `playwright`/`chrome-devtools`: To identify UI "jank" or DOM complexity.
-
-## Agent-Specific Simplification Protocols
-
-### [Senior Backend Engineer](file:///.opencode/agents/backend.md)
-- **State Reduction**: Derive values instead of storing them. If `is_active` can be derived from `status === 'active'`, delete the redundant field.
-- **Query Optimization**: Use `sqlite` to check for N+1 issues. Replace nested loops with single optimized queries or batching.
-- **Error Flattening**: Centralize error handling middleware. Remove repetitive `try/catch` blocks from business logic to expose the "Happy Path".
-
-### [Senior Frontend Engineer](file:///.opencode/agents/frontend.md)
-- **DOM Streamlining**: Replace "Div Soup" with semantic HTML elements.
-- **Hook Optimization**: Audit `useEffect` usage. If it's used to sync state, replace it with derived constants or `useMemo`.
-- **Styling Discipline**: Replace custom inline styles with standard Tailwind utility classes or `cva` variants.
-
-### [QA Automation Engineer](file:///.opencode/agents/quality.md)
-- **Test Pruning**: Remove redundant unit tests if the logic is covered by a high-confidence integration test.
-- **Selector Simplification**: Replace fragile XPaths or class-based selectors with accessible `getByRole` or `getByLabel` locators.
-
-## The Procedural Workflow
-
-### 1. The Deletion Test
-- **Action**: Use `grep_search` to find references to the target code. If references < 1, **DELETE**.
-- **Rule**: Dead code is the best code.
-
-### 2. The Library Leverage (via `context7`)
-- **Action**: Use `context7` to check if a complex custom utility (e.g., date formatting, deep cloning) can be replaced by a native function or a lightweight standard library (Lodash/Zustand).
-
-### 3. Flat Logic Audit (via `sequential-thinking`)
-- **Action**: Step through the logic. If nesting > 3 levels or cyclomatic complexity is high, apply guard clauses.
-- **Target Pattern**:
-  ```diff
-  - if (user) {
-  -   if (user.active) {
-  -     return doThing();
-  -   }
-  - }
-  + if (!user?.active) return;
-  + return doThing();
-  ```
-
-## Complexity Metrics to Watch
-- **Nesting Level**: Maximum 3 levels of indentation.
-- **Function Size**: Maximum 30 lines.
-- **Cognitive Weight**: If a function requires `sequential-thinking` just to "read" it, it's too complex.
-
-## Output Format
-
-```markdown
-## Simplification Review: [Target File/Function]
-
-### 1. Opportunities Identified
-| Area | Description | Tool Used | Impact |
-|------|-------------|-----------|--------|
-| Logic | Flatten nested `if` statements | `sequential-thinking` | High |
-| State | Remove redundant `isLoading` boolean | - | Medium |
-| Lib | Replace custom deep clone with `structuredClone` | `context7` | Low |
-
-### 2. Implementation: Before vs After
-**Before**:
-```javascript
-// Complex, nested, or redundant code
-```
-
-**After**:
-```javascript
-// Streamlined, readable, and efficient code
-```
-
-### 3. Impact Assessment
-- Lines Removed: [Count]
-- Performance Gain: [e.g., Reduced O(N²) to O(N)]
-- Maintenance Status: Optimized
-```
+---
+name: simplification
+description: Complexity reduction and architectural streamlining for Frontend, Backend, and Quality agents.
+---
+
+# Simplification Skill (Occam's Razor)
+
+## Purpose
+"Simplicity is the ultimate sophistication." This skill provides a rigorous protocol for identifying and eliminating over-engineering, code bloat, and technical debt across the entire stack.
+
+## Overview
+The Simplification Skill is a cross-functional protocol used to flatten logic, reduce state, and ensure every line of code justifies its existence. It leverages specialized agents and MCP tools to validate architectural simplicity and performance.
+
+## Prerequisites
+- **Agents**: [Backend](file:///.opencode/agents/backend.md), [Frontend](file:///.opencode/agents/frontend.md), [Quality](file:///.opencode/agents/quality.md).
+- **MCP Tools**: 
+    - `context7`: To research lightweight alternatives and library standards.
+    - `sequential-thinking`: To audit complex logic flows for potential flattening.
+    - `sqlite`: To verify if database schemas or queries can be compressed.
+    - `playwright`/`chrome-devtools`: To identify UI "jank" or DOM complexity.
+
+## Agent-Specific Simplification Protocols
+
+### [Senior Backend Engineer](file:///.opencode/agents/backend.md)
+- **State Reduction**: Derive values instead of storing them. If `is_active` can be derived from `status === 'active'`, delete the redundant field.
+- **Query Optimization**: Use `sqlite` to check for N+1 issues. Replace nested loops with single optimized queries or batching.
+- **Error Flattening**: Centralize error handling middleware. Remove repetitive `try/catch` blocks from business logic to expose the "Happy Path".
+
+### [Senior Frontend Engineer](file:///.opencode/agents/frontend.md)
+- **DOM Streamlining**: Replace "Div Soup" with semantic HTML elements.
+- **Hook Optimization**: Audit `useEffect` usage. If it's used to sync state, replace it with derived constants or `useMemo`.
+- **Styling Discipline**: Replace custom inline styles with standard Tailwind utility classes or `cva` variants.
+
+### [QA Automation Engineer](file:///.opencode/agents/quality.md)
+- **Test Pruning**: Remove redundant unit tests if the logic is covered by a high-confidence integration test.
+- **Selector Simplification**: Replace fragile XPaths or class-based selectors with accessible `getByRole` or `getByLabel` locators.
+
+## The Procedural Workflow
+
+### 1. The Deletion Test
+- **Action**: Use `grep_search` to find references to the target code. If references < 1, **DELETE**.
+- **Rule**: Dead code is the best code.
+
+### 2. The Library Leverage (via `context7`)
+- **Action**: Use `context7` to check if a complex custom utility (e.g., date formatting, deep cloning) can be replaced by a native function or a lightweight standard library (Lodash/Zustand).
+
+### 3. Flat Logic Audit (via `sequential-thinking`)
+- **Action**: Step through the logic. If nesting > 3 levels or cyclomatic complexity is high, apply guard clauses.
+- **Target Pattern**:
+  ```diff
+  - if (user) {
+  -   if (user.active) {
+  -     return doThing();
+  -   }
+  - }
+  + if (!user?.active) return;
+  + return doThing();
+  ```
+
+## Complexity Metrics to Watch
+- **Nesting Level**: Maximum 3 levels of indentation.
+- **Function Size**: Maximum 30 lines.
+- **Cognitive Weight**: If a function requires `sequential-thinking` just to "read" it, it's too complex.
+
+## Output Format
+
+```markdown
+## Simplification Review: [Target File/Function]
+
+### 1. Opportunities Identified
+| Area | Description | Tool Used | Impact |
+|------|-------------|-----------|--------|
+| Logic | Flatten nested `if` statements | `sequential-thinking` | High |
+| State | Remove redundant `isLoading` boolean | - | Medium |
+| Lib | Replace custom deep clone with `structuredClone` | `context7` | Low |
+
+### 2. Implementation: Before vs After
+**Before**:
+```javascript
+// Complex, nested, or redundant code
+```
+
+**After**:
+```javascript
+// Streamlined, readable, and efficient code
+```
+
+### 3. Impact Assessment
+- Lines Removed: [Count]
+- Performance Gain: [e.g., Reduced O(N²) to O(N)]
+- Maintenance Status: Optimized
+```