@hienlh/ppm 0.9.81 → 0.9.82

package/.opencode/skills/google-adk-python/references/sessions-state-memory-artifacts.md

@@ -1,131 +0,0 @@
-# Sessions, State, Memory, Artifacts
-
-## ToolContext.state
-
-Ephemeral dict persisting across tool calls within session.
-
-```python
-from google.adk.tools.tool_context import ToolContext
-
-def save_user(user_id: str, name: str, tool_context: ToolContext) -> dict:
-    if "users" not in tool_context.state:
-        tool_context.state["users"] = {}
-    tool_context.state["users"][user_id] = name
-    return {"status": "saved"}
-
-def get_user(user_id: str, tool_context: ToolContext) -> str:
-    return tool_context.state.get("users", {}).get(user_id, "not_found")
-
-def increment(tool_context: ToolContext) -> int:
-    count = tool_context.state.get("counter", 0) + 1
-    tool_context.state["counter"] = count
-    return count
-```
-
-Lifecycle: Created per session, cleared on end, not shared.
-
-## Artifacts
-
-Versioned binary storage.
-
-```python
-from google.genai import types
-from google.adk.tools.tool_context import ToolContext
-
-async def save_text(name: str, content: str, tool_context: ToolContext) -> dict:
-    part = types.Part(inline_data=types.Blob(mime_type="text/plain", data=content.encode()))
-    version = await tool_context.save_artifact(name, part)
-    return {"name": name, "version": version}
-
-async def save_image(name: str, bytes: bytes, tool_context: ToolContext) -> dict:
-    part = types.Part(inline_data=types.Blob(mime_type="image/png", data=bytes))
-    version = await tool_context.save_artifact(name, part)
-    return {"name": name, "version": version}
-
-async def load_text(name: str, tool_context: ToolContext) -> str:
-    artifact = await tool_context.load_artifact(name)
-    return artifact.inline_data.data.decode() if artifact else None
-
-async def load_version(name: str, version: int, tool_context: ToolContext) -> str:
-    artifact = await tool_context.load_artifact(name, version=version)
-    return artifact.inline_data.data.decode() if artifact else None
-```
-
-Versioning: Each save creates new version (1, 2, 3...), load without version returns latest.
-
-## Session Services
-
-```python
-from google.adk.sessions.in_memory_session_service import InMemorySessionService
-from google.adk.sessions.vertex_ai_session_service import VertexAiSessionService
-from google.adk.sessions.spanner_session_service import SpannerSessionService
-from google.adk.runners import Runner
-
-# Dev
-session_service = InMemorySessionService()
-
-# Prod (Vertex)
-session_service = VertexAiSessionService(project_id="my-project", location="us-central1")
-
-# Prod (Spanner)
-session_service = SpannerSessionService(
-    project_id="my-project", instance_id="my-instance", database_id="sessions"
-)
-
-runner = Runner(agent=root_agent, session_service=session_service)
-```
-
-## Memory
-
-Long-term recall across sessions.
-
-```python
-from google.adk.memory import MemoryService
-
-memory_service = MemoryService(project_id="my-project", location="us-central1")
-runner = Runner(agent=root_agent, session_service=session_service, memory_service=memory_service)
-```
-
-**Comparison:** State (ephemeral, dict), Artifacts (versioned binary, session), Memory (long-term, cross-session)
-
-## Runner
-
-```python
-from google.adk.runners import Runner
-
-runner = Runner(agent=root_agent, session_service=session_service)
-
-# Production (streaming)
-async def execute():
-    async for event in runner.run_async("Query", session_id="user-123"):
-        if event.text: print(event.text, end="")
-        if event.tool_call: print(f"\nTool: {event.tool_call.name}")
-
-# Bidi-streaming
-async def live():
-    async for event in runner.run_live(session_id="user-456"):
-        if event.server_content: print(event.server_content)
-        if user_has_input(): await event.send(get_user_input())
-
-# Sync (debug)
-result = runner.run("Query", session_id="debug")
-```
-
-## Invocation Lifecycle
-
-1. Session retrieval 2. Context creation 3. Agent execution 4. Event streaming 5. Compaction
-
-```python
-runner = Runner(agent=root_agent, session_service=session_service, compaction_threshold=20)
-await runner.compact_session("long-session")
-```
-
-## Best Practices
-
-- State: temporary session data
-- Artifacts: binary/large data
-- Memory: long-term knowledge
-- Session service by env (InMemory dev, Vertex/Spanner prod)
-- Enable compaction for long conversations
-- run_async for production, run_live for bidirectional, run for debug
-- Clean old sessions, version artifacts

package/.opencode/skills/google-adk-python/references/tools-and-mcp-integration.md

@@ -1,146 +0,0 @@
-# Tools and MCP Integration
-
-## Custom Tools
-
-Python functions (sync/async) with optional `ToolContext`.
-
-```python
-def calculator(operation: str, a: float, b: float) -> float:
-    """Perform math operations.
-    Args: operation ('add'|'subtract'|'multiply'|'divide'), a, b
-    Returns: Result
-    """
-    ops = {"add": a + b, "subtract": a - b, "multiply": a * b, "divide": a / b}
-    return ops[operation]
-```
-
-**Async with ToolContext:**
-
-```python
-from google.adk.tools.tool_context import ToolContext
-
-async def fetch(url: str, tool_context: ToolContext) -> dict:
-    cache = tool_context.state.get("cache", {})
-    if url in cache: return cache[url]
-    data = await http_client.get(url)
-    cache[url] = data
-    tool_context.state["cache"] = cache
-    return data
-```
-
-## ToolContext
-
-Props: `state` (dict), `save_artifact()`, `load_artifact()`
-
-```python
-def store(key: str, value: str, tool_context: ToolContext) -> dict:
-    tool_context.state[key] = value
-    return {"status": "saved"}
-
-from google.genai import types
-
-async def save_doc(name: str, content: str, tool_context: ToolContext) -> dict:
-    part = types.Part(inline_data=types.Blob(mime_type="text/plain", data=content.encode()))
-    version = await tool_context.save_artifact(name, part)
-    return {"name": name, "version": version}
-
-async def load_doc(name: str, tool_context: ToolContext) -> str:
-    artifact = await tool_context.load_artifact(name)
-    return artifact.inline_data.data.decode() if artifact else None
-```
-
-## Built-in Tools
-
-```python
-from google.adk.tools import google_search, load_artifacts, BuiltInCodeExecutor
-
-executor = BuiltInCodeExecutor()
-agent = Agent(name="agent", model="gemini-2.5-flash", tools=[google_search, load_artifacts, executor])
-```
-
-## MCP Integration
-
-```python
-from google.adk.tools.mcp_tool import StdioConnectionParams
-from google.adk.tools.mcp_tool.mcp_toolset import MCPToolset
-from mcp import StdioServerParameters
-
-filesystem_mcp = MCPToolset(
-    connection_params=StdioConnectionParams(
-        server_params=StdioServerParameters(
-            command='npx',
-            args=['-y', '@modelcontextprotocol/server-filesystem', '/home/user/docs']
-        ),
-        timeout=5
-    ),
-    tool_filter=['read_file', 'list_directory']
-)
-
-agent = Agent(name="agent", model="gemini-2.5-flash", tools=[filesystem_mcp])
-```
-
-**Multiple servers:**
-
-```python
-git_mcp = MCPToolset(
-    connection_params=StdioConnectionParams(
-        server_params=StdioServerParameters(command='npx', args=['-y', '@modelcontextprotocol/server-git']),
-        timeout=10
-    )
-)
-
-postgres_mcp = MCPToolset(
-    connection_params=StdioConnectionParams(
-        server_params=StdioServerParameters(
-            command='npx', args=['-y', '@modelcontextprotocol/server-postgres', 'postgresql://localhost/db']
-        ),
-        timeout=15
-    )
-)
-
-agent = Agent(name="agent", model="gemini-2.5-flash", tools=[filesystem_mcp, git_mcp, postgres_mcp])
-```
-
-## Tool Filtering
-
-```python
-# List
-MCPToolset(connection_params=params, tool_filter=['read_file', 'write_file'])
-
-# Lambda
-MCPToolset(connection_params=params, tool_filter=lambda t: t.name.startswith('read_'))
-```
-
-## LongRunningFunctionTool
-
-```python
-from google.adk.tools.long_running_tool import LongRunningFunctionTool
-
-async def approve(amount: float) -> dict:
-    return {"approved": True}
-
-tool = LongRunningFunctionTool(func=approve)
-agent = Agent(name="agent", model="gemini-2.5-flash", tools=[tool])
-```
-
-## ExampleTool
-
-```python
-from google.adk.tools.example_tool import ExampleTool
-
-examples = ExampleTool(examples=[
-    {"input": "2+2?", "output": "4"},
-    {"input": "Capital of France?", "output": "Paris"}
-])
-
-agent = Agent(name="agent", model="gemini-2.5-flash", tools=[examples])
-```
-
-## Best Practices
-
-- Docstrings with Args/Returns for schema
-- Async for I/O
-- ToolContext.state for session data
-- Artifacts for binary/large data
-- Filter MCP tools to reduce context
-- Validate inputs

package/.opencode/skills/hub/SKILL.md

@@ -1,55 +0,0 @@
----
-name: ckm:hub
-description: Open Content Hub + Marketing Dashboard
-argument-hint: "[--stop|--scan]"
-metadata:
-  author: claudekit
-  version: "1.0.0"
----
-
-💡
-Activate `content-hub` skill and start Marketing Dashboard.
-
-## Usage
-
-```bash
-/write:hub           # Start all services
-/write:hub --scan    # Rescan assets folder
-/write:hub --stop    # Stop all servers
-```
-
-## Workflow
-
-1. **Start servers**: Content Hub + Marketing Dashboard (API + Frontend)
-2. **Auto-open**: Content Hub at http://localhost:3457/hub
-3. **Access dashboard**: Marketing Dashboard at http://localhost:5173
-
-## Services Started
-
-| Service | URL | Description |
-|---------|-----|-------------|
-| Content Hub | http://localhost:3457/hub | Asset gallery with AI editor |
-| Dashboard UI | http://localhost:5173 | Marketing dashboard (Vue) |
-| Dashboard API | http://localhost:3457/api/ | REST API (Hono + SQLite) |
-
-## Execution
-
-```bash
-bash .opencode/skills/content-hub/scripts/start-all.sh $ARGUMENTS
-```
-
-Report all URLs when servers start.
-
-## Features
-
-**Content Hub:**
-- Visual asset gallery with thumbnails
-- AI-powered content editor
-- Brand context sidebar
-- Filter/search by type
-
-**Marketing Dashboard:**
-- Campaign management
-- Content library
-- Asset linking
-- Automation panel

package/.opencode/skills/init/SKILL.md

@@ -1,219 +0,0 @@
----
-name: ckm:init
-description: 💡💡💡💡 Initialize marketing project
-argument-hint: "[prompt]"
-metadata:
-  author: claudekit
-  version: "1.0.0"
----
-
-Initialize a new marketing project by gathering comprehensive client information through an agency-style marketing brief interview.
-
-## Your Mission
-
-Conduct a structured interview to gather all essential marketing information, then generate foundational project documentation.
-
-<user-input>$ARGUMENTS</user-input>
-
-**IMPORTANT**:
-- Do not skip any phase or section.
-- Analyze the skills catalog and activate the skills that are needed for the task during the process.
-- Ensure token efficiency while maintaining high quality.
-- Sacrifice grammar for the sake of concision when writing reports.
-- In reports, list any unresolved questions at the end, if any.
-
-## Phase 0: Git Setup
-
-1. **Check Git Status**
-   - Run `git status` to detect if git is initialized
-   - If not a git repo, ask user: "Initialize git repository?"
-
-2. **If User Confirms Git Init**
-   - Run `git init`
-   - Ask: "Create remote repository?" (GitHub/GitLab/etc.)
-   - If yes, gather repo name and create via `gh repo create`
-
-3. **If Already Initialized**
-   - Show current remote (if any)
-   - Proceed to Phase 1
-
-## Phase 1: Quick Evaluation
-
-**Main agent spawns parallel researchers** to evaluate current project situation:
-
-### 1.1 Parallel Research (spawn 3-4 `researcher` agents simultaneously)
-
-| Agent | Focus Area | Data Sources |
-|-------|------------|--------------|
-| Researcher 1 | Product/Service Analysis | `README.md`, `CLAUDE.md`, `package.json`, codebase |
-| Researcher 2 | Existing Marketing Assets | `docs/*.md`, `assets/`, marketing collaterals |
-| Researcher 3 | Digital Presence | Website URL (if found), social links, SEO status |
-| Researcher 4 | Competitive Landscape | Industry context, market positioning |
-
-### 1.2 If Docs Exist in `docs/`
-
-1. **Read all existing docs**:
-   - `docs/project-overview-pdr.md`
-   - `docs/marketing-overview.md`
-   - `docs/brand-guidelines.md`
-   - `docs/design-guidelines.md`
-   - Any other `docs/*.md` files
-
-2. **Compile findings summary** with key points:
-   - Product/service description
-   - Target audience (if defined)
-   - Brand identity (if defined)
-   - Marketing objectives (if defined)
-   - Existing campaigns/assets
-
-3. **Verify with user via `AskUserQuestion`**:
-   - Display compiled summary
-   - Ask: "Is this information still accurate?"
-   - Options: "Yes, proceed" / "Needs updates" / "Start fresh"
-   - If "Needs updates": ask which sections need revision
-
-### 1.3 If No Docs Exist
-
-- Note: "No existing docs found"
-- Proceed directly to Phase 2
-
-## Phase 2: Marketing Brief Interview
-
-Use `AskUserQuestion` tool for each section.
-
-### Section A: Basic Information
-- Company/Product name and tagline
-- Industry/Niche
-- Brief introduction (2-3 sentences)
-- Website URL (if exists): ask for URL if not found, read and analyze website to extract information
-
-### Section B: Products/Services
-- What are your main products/services?
-- Key USPs and differentiators
-- Price range/Pricing model
-- Top 3 product highlights
-
-### Section C: Objectives (multi-select)
-- Brand awareness
-- Lead generation
-- Sales conversion
-- Customer retention
-- Market expansion
-- Product launch
-- Other
-
-### Section D: Target Audience
-- **Geographic**: Countries/Regions
-- **Demographic**: Age, gender, income level
-- **Psychographic**: Interests, values, lifestyle
-- **B2B specifics**: Company size, industry, decision-maker roles
-
-### Section E: Competitive Landscape
-- Top 3 competitors (names/URLs)
-- Your competitive advantage
-- Market positioning goal
-
-### Section F: Brand Assets
-- Logo available? (Y/N, format)
-- Brand colors defined?
-- Brand fonts specified?
-- Existing brand guidelines doc?
-- Marketing collaterals (name cards, catalogue, etc.)?
-
-### Section G: Design Guidelines
-Ask if they want to extract from:
-- Website (provide URL)
-- Videos (provide links)
-- Existing documents (provide paths)
-- Manual input
-
-### Section H: Social Presence
-- Active social platforms (multi-select)
-- Social profile URLs
-- Current follower counts (approximate)
-
-### Section I: Timeline & Budget
-- Expected launch/go-live date
-- Monthly marketing budget range
-- Priority channels for budget allocation
-
-## Phase 3: Documentation Generation
-
-**IMPORANT**: Create/update following documents even docs exists in Step 1.2
-
-After interview completion, spawn `docs-manager` agent with gathered data to:
-
-1. **Update `CLAUDE.md`**
-   - Add marketing context section
-   - Update role description
-
-2. **Create/Update `README.md`**
-   - Project overview
-   - Quick start for marketing workflows
-
-3. **Generate `docs/project-overview-pdr.md`**
-   - Full product development requirements
-   - Marketing objectives and KPIs
-   - Target audience profiles
-   - Brand guidelines summary
-
-4. **Generate `docs/marketing-overview.md`**
-   - Marketing strategy summary
-   - Channel priorities
-   - Campaign roadmap outline
-
-5. **Generate `docs/project-roadmap.md`**
-   - Marketing milestones
-   - Campaign timeline
-   - Resource allocation
-
-## Phase 3.5: Commit Changes
-
-Ask user: "Commit generated documentation?"
-- If yes, execute `/git:cm` command to stage and commit
-- Suggested commit message: "docs: initialize marketing project documentation"
-
-## Phase 4: Next Steps
-
-After documentation generated:
-
-1. **Analyze Available Commands**
-   - Run `python .opencode/scripts/generate_catalogs.py --commands`
-   - Identify relevant next steps
-
-2. **Suggest Next Actions** (pick 3-5 most relevant):
-   - `/persona create` - Create customer personas
-   - `/seo audit` - Audit website SEO
-   - `/funnel design` - Design marketing funnel
-   - `/campaign create` - Plan first campaign
-   - `/social schedule` - Create content calendar
-   - `/write:blog` - Start blog content
-   - `/competitor` - Deep competitor analysis
-   - None - finish the process.
-
-## Agents Used
-- `researcher` (x4 parallel) - Quick evaluation of project situation
-- `docs-manager` - Documentation generation
-
-## Skills Used
-- `research` - Quick evaluation of project situation
-- `chrome-devtools` - Website extraction
-- `ai-multimodal` - Video/image analysis
-- `assets-organizing` - Standardized output paths and naming conventions
-
-## Commands Used
-- `/git:cm` - Stage and commit changes
-
-## Output
-**IMPORANT**: Create/update following documents even docs exists in Step 1.2
-- `CLAUDE.md` - Updated with marketing context
-- `README.md` - Project overview
-- `docs/project-overview-pdr.md` - Full PDR
-- `docs/marketing-overview.md` - Marketing strategy
-- `docs/project-roadmap.md` - Campaign timeline
-
-## Notes
-- Use progressive disclosure - don't overwhelm with all questions at once
-- Validate responses against existing docs if available
-- If user skips optional sections, note as "TBD" in docs
-- Keep interview conversational, not interrogative

package/.opencode/skills/journal/SKILL.md

@@ -1,22 +0,0 @@
----
-name: ck:journal
-description: 💡 Write some journal entries.
-metadata:
-  author: claudekit
-  version: "1.0.0"
----
-
-## Writing Style Detection
-
-Before writing, check for a writing-styles directory:
-1. Check `docs/writing-styles/` — if exists, read all `.md` files inside
-2. Check `assets/writing-styles/` — if exists and step 1 not found, read all `.md` files inside
-3. If writing styles found → adopt tone, vocabulary, sentence structure, and formatting rules from those files
-4. If no writing-styles directory found → write freely in a natural, conversational tone
-
-## Journal Writing
-
-Use the `journal-writer` subagent to explore the memories and recent code changes, and write journal entries.
-- Concise, focused on important events, key changes, impacts, and decisions
-- Apply detected writing style (if any) to the journal voice
-- Keep journal entries in `docs/journal/` directory

package/.opencode/skills/kanban/SKILL.md

@@ -1,69 +0,0 @@
----
-name: ck:kanban
-description: AI agent orchestration board (Coming Soon)
-metadata:
-  author: claudekit
-  version: "1.0.0"
----
-
-Plans dashboard with progress tracking and timeline visualization.
-
-## Usage
-
-- `/kanban` - View dashboard for ./plans directory
-- `/kanban plans/` - View dashboard for specific directory
-- `/kanban --stop` - Stop running server
-
-## Features
-
-- Plan cards with progress bars
-- Phase status breakdown (completed, in-progress, pending)
-- Timeline/Gantt visualization
-- Activity heatmap
-- Issue and branch links
-
-## Execution
-
-```bash
-# Check if --stop flag
-if [[ "$ARGUMENTS" == *"--stop"* ]]; then
-  node .opencode/skills/plans-kanban/scripts/server.cjs --stop
-  exit 0
-fi
-
-# Determine plans directory
-INPUT_DIR="{{dir}}"
-PLANS_DIR="${INPUT_DIR:-./plans}"
-
-# Start kanban dashboard
-node .opencode/skills/plans-kanban/scripts/server.cjs \
-  --dir "$PLANS_DIR" \
-  --host 0.0.0.0 \
-  --open \
-  --background
-```
-
-Report the URL to the user. For remote access from other devices, use the `networkUrl` field from the JSON output.
-
-## Future Plans
-
-The `/kanban` command will evolve into **VibeKanban-inspired** AI agent orchestration:
-
-### Phase 1 (Current - MVP)
-- ✅ Task board with progress tracking
-- ✅ Visual representation of plans/tasks
-- ✅ Click to view plan details
-
-### Phase 2 (Worktree Integration)
-- Create tasks → spawn git worktrees
-- Assign agents to tasks
-- Track agent progress per worktree
-
-### Phase 3 (Full Orchestration)
-- Parallel agent execution monitoring
-- Code diff/review interface
-- PR creation workflow
-- Agent output streaming
-- Conflict detection
-
-Track progress: https://github.com/claudekit/claudekit-engineer/issues/189