Merge pull request #21 from NeoLabHQ/feature/vg/add-new-context-engineering-commands

feat: Add FPF plugin and LLM-as-Judge based commands for SADD

Author: LeoVS09

.beads/issues/cek-31ce.md

@@ -1,7 +1,7 @@
 {
   "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
   "name": "context-engineering-kit",
-  "version": "1.5.0",
+  "version": "1.6.0",
   "description": "Hand-crafted collection of advanced context engineering techniques and patterns with minimal token footprint focused on improving agent result quality.",
   "owner": {
     "name": "NeoLabHQ",
@@ -11,7 +11,7 @@
     {
       "name": "reflexion",
       "description": "Collection of commands that force LLM to reflect on previous response and output. Based on papers like Self-Refine and Reflexion. These techniques improve the output of large language models by introducing feedback and refinement loops.",
-      "version": "1.1.0",
+      "version": "1.1.2",
       "author": {
         "name": "Vlad Goncharov",
         "email": "[email protected]"
@@ -33,7 +33,7 @@
     {
       "name": "git",
       "description": "Introduces commands for commit and PRs creation.",
-      "version": "1.0.1",
+      "version": "1.0.2",
       "author": {
         "name": "Vlad Goncharov",
         "email": "[email protected]"
@@ -55,7 +55,7 @@
     {
       "name": "sadd",
       "description": "Introduces skills for subagent-driven development, dispatches fresh subagent for each task with code review between tasks, enabling fast iteration with quality gates.",
-      "version": "1.0.0",
+      "version": "1.1.0",
       "author": {
         "name": "Vlad Goncharov",
         "email": "[email protected]"
@@ -99,7 +99,7 @@
     {
       "name": "customaize-agent",
       "description": "Commands and skills for writing and refining commands, hooks, skills for Claude Code, includes Anthropic Best Practices and Agent Persuasion Principles that can be useful for sub-agent workflows.",
-      "version": "1.1.0",
+      "version": "1.2.0",
       "author": {
         "name": "Vlad Goncharov",
         "email": "[email protected]"
@@ -139,6 +139,17 @@
       },
       "source": "./plugins/mcp",
       "category": "development"
+    },
+    {
+      "name": "fpf",
+      "description": "First Principles Framework (FPF) for structured reasoning. Implements ADI (Abduction-Deduction-Induction) cycle for hypothesis generation, logical verification, empirical validation, and auditable decision-making.",
+      "version": "1.0.0",
+      "author": {
+        "name": "Vlad Goncharov",
+        "email": "[email protected]"
+      },
+      "source": "./plugins/fpf",
+      "category": "development"
     }
   ]
 }

.claude-plugin/marketplace.json

@@ -0,0 +1,138 @@
+---
+description: "Create a minibeads task with proper structure and context. User prompt: 'create /add-task command in @.claude/commands/ based on mb'"
+argument-hint: Task title or description (e.g., "Add validation to form inputs")
+allowed-tools: Bash(mb create:*), Bash(mb show:*), Bash(mb list:*), Bash(mb dep:*), Bash(mb ready:*), Skill(refine-task)
+---
+
+# Add Task Command
+
+<task>
+You are a task creation specialist. Create well-structured minibeads tasks that are clear, actionable, and properly categorized.
+</task>
+
+<context>
+This command helps create tasks using minibeads (mb) with:
+
+1. Clear, action-oriented titles
+2. Appropriate type classification
+3. Useful description when needed
+4. Automatic refinement
+</context>
+
+<workflow>
+
+## Phase 1: Analyze Input
+
+1. **Parse the user's request**:
+   - Extract the core task objective
+   - Identify implied type (bug, feature, task)
+
+2. **Clarify if ambiguous**:
+   - Is this a bug fix or new feature?
+   - Any related tasks or dependencies?
+
+## Phase 2: Structure the Task
+
+1. **Create action-oriented title**:
+   - Start with verb: Add, Fix, Update, Implement, Remove, Refactor
+   - Be specific but concise
+   - Examples:
+     - "Add validation to login form"
+     - "Fix null pointer in user service"
+     - "Implement caching for API responses"
+
+2. **Determine type**:
+   | Type | Use When |
+   |------|----------|
+   | `task` | General work items, refactoring, maintenance |
+   | `bug` | Something is broken or not working correctly |
+   | `feature` | New functionality or capability |
+
+## Phase 3: Create Task
+
+```bash
+mb create "TITLE" -t TYPE -d "initial user prompt: [EXACT USER PROMPT]
+
+[Short description of what the task involves]"
+```
+
+The description (-d) MUST include:
+1. **Line 1**: `initial user prompt: ` followed by the EXACT user input as provided
+2. **Line 2+**: Short description of the task scope and context
+
+## Phase 4: Refine Task
+
+After creating the task, **ALWAYS** invoke the refine-task skill to add implementation details:
+
+```
+Skill tool invocation:
+- skill: "refine-task"
+- args: "<task-id>"
+```
+
+This ensures every created task has:
+- Detailed specification
+- Affected files identified
+- Implementation resources gathered
+
+**IMPORTANT**: Do not skip this step. The refinement is mandatory for all created tasks.
+
+</workflow>
+
+<examples>
+
+**Simple task:**
+```bash
+mb create "Add unit tests for auth module" -t task -d "initial user prompt: add tests for auth
+
+Cover login, logout, and session management functions."
+```
+
+**Bug with context:**
+```bash
+mb create "Fix login timeout on slow connections" -t bug -d "initial user prompt: users getting 504 errors on slow wifi
+
+Increase timeout threshold and add retry logic for network failures."
+```
+
+**Feature request:**
+```bash
+mb create "Implement dark mode toggle" -t feature -d "initial user prompt: add dark mode to settings page
+
+Add theme switching in settings with localStorage persistence."
+```
+
+</examples>
+
+<multiple_tasks>
+When creating multiple related tasks:
+
+1. Create tasks in parallel when possible
+2. Add dependencies after creation:
+   ```bash
+   mb dep add <dependent-task> <dependency-task>
+   ```
+
+Example flow:
+```bash
+mb create "Implement user API" -t feature -d "initial user prompt: need user management endpoints
+
+REST endpoints for user CRUD operations."
+
+mb create "Add user API tests" -t task -d "initial user prompt: need user management endpoints
+
+Unit and integration tests for user API."
+
+mb dep add cek-tests cek-api    # Tests depend on API
+```
+</multiple_tasks>
+
+<output>
+After creating and refining the task:
+
+1. **Show created task**: `mb show <task-id>`
+2. **Confirm refinement completed**
+3. **Suggest next steps**:
+   - Add dependencies if related tasks exist
+   - Start work with `mb update <task-id> --status=in_progress`
+</output>