Initial commit

2025-11-30 09:04:17 +08:00
commit 652915e226
62 changed files with 8212 additions and 0 deletions
--- a/commands/agents-create.md
+++ b/commands/agents-create.md
@@ -0,0 +1,216 @@
+---
+description: Create agent(s) using templates - supports parallel creation for 3+ agents
+argument-hint: <agent-name> "<description>" [--marketplace] | <agent-1> "<desc-1>" <agent-2> "<desc-2>" ... [--marketplace]
+allowed-tools: Task, Read, Write, Edit, Bash, Glob, Grep, TodoWrite
+---
+
+---
+🚨 **EXECUTION NOTICE FOR CLAUDE**
+
+When you invoke this command via `SlashCommand(/domain-plugin-builder:agents-create ...)`, the system returns THESE INSTRUCTIONS below.
+
+**YOU are the executor. This is NOT an autonomous subprocess.**
+
+- ✅ The phases below are YOUR execution checklist
+- ✅ YOU must run each phase immediately using tools (Bash, Read, Write, Edit, TodoWrite)
+- ✅ Complete ALL phases before considering this command done
+- ❌ DON'T wait for "the command to complete" - YOU complete it by executing the phases
+- ❌ DON'T treat this as status output - it IS your instruction set
+
+**Immediately after SlashCommand returns, start executing Phase 0, then Phase 1, etc.**
+
+See `@CLAUDE.md` section "SlashCommand Execution - YOU Are The Executor" for detailed explanation.
+
+---
+
+**Arguments**: $ARGUMENTS
+
+**Security**: Follow @~/.claude/plugins/marketplaces/dev-lifecycle-marketplace/docs/security/SECURITY-RULES.md (never hardcode API keys)
+
+**Framework**: Load @component-decision-framework.md and @dans-composition-pattern.md
+
+**Goal**: Create properly structured agent file(s). For 3+ agents, use parallel execution.
+
+**EXECUTE each phase immediately** using actual tools (Bash, Read, Write, Task, TodoWrite). Complete all phases in order.
+
+Phase 0: Create Todo List
+
+Create todo list for all phases below.
+
+Phase 1: Parse Arguments, Determine Mode & Count Agents
+Goal: Extract agent specifications, determine mode, and execution strategy
+
+Actions:
+
+Parse $ARGUMENTS to extract:
+- Agent names and descriptions
+- Plugin name (from context or pwd)
+- Marketplace mode (check for --marketplace flag)
+
+If plugin not specified:
+
+!{bash basename $(pwd)}
+
+Determine base path - check if already in a plugin directory:
+
+!{bash test -f .claude-plugin/plugin.json && echo "." || (echo "$ARGUMENTS" | grep -q "\-\-marketplace" && echo "plugins/$(basename $(pwd))" || echo ".")}
+
+Store as $BASE_PATH:
+- If .claude-plugin/plugin.json exists: BASE_PATH="." (already in plugin directory)
+- Else if --marketplace present: BASE_PATH="plugins/$PLUGIN_NAME"
+- Else: BASE_PATH="." (standalone plugin mode)
+
+All subsequent file operations use $BASE_PATH instead of hardcoded "plugins/$PLUGIN_NAME"
+
+Use bash to parse $ARGUMENTS and count how many agents are being requested:
+
+!{bash echo "$ARGUMENTS" | grep -oE '<[^>]+>' | wc -l}
+
+Store the count. Then extract each agent specification:
+- If count = 1: Single agent mode - extract <agent-name> and "<description>"
+- If count = 2: Two agents mode - extract both <agent-N> "<desc-N>" sets
+- If count >= 3: Multiple agents mode - extract all <agent-N> "<desc-N>" sets
+
+All agents use Task tool - whether creating 1 or 10 agents.
+
+**Note:** Agents inherit tools from parent - no need to specify tools field.
+
+Phase 2: Load Templates
+Goal: Study framework patterns
+
+Actions:
+
+Load agent template immediately:
+!{Read @agent-with-phased-webfetch.md}
+
+Determine plugin location from context (default: domain-plugin-builder)
+
+Phase 3: Create Agent(s)
+Goal: Generate agent file(s) efficiently
+
+Actions:
+
+**Decision: 1-2 agents = build directly, 3+ agents = use Task() for parallel**
+
+**For 1-2 Agents:**
+
+Build directly - execute these steps immediately:
+
+1. Load template:
+!{Read @agent-with-phased-webfetch.md}
+
+2. Load color decision framework:
+!{Read @agent-color-decision.md}
+
+3. For each agent, create the file:
+!{Write $BASE_PATH/agents/$AGENT_NAME.md}
+
+Include:
+- Frontmatter with name, description, model: inherit, color (from decision framework)
+- **NO tools field** - agents inherit tools from parent
+- "Available Tools & Resources" section specifying:
+  - MCP servers (e.g., mcp__github, mcp__supabase)
+  - Skills (e.g., Skill(plugin:skill-name))
+  - Slash commands (e.g., SlashCommand(/plugin:command-name))
+- Progressive WebFetch for documentation
+- Keep under 300 lines
+
+4. Validate:
+!{Bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-agent.sh $BASE_PATH/agents/$AGENT_NAME.md}
+
+No need for Task() overhead when building 1-2 agents
+
+**For 3+ Agents:**
+
+**CRITICAL: Send ALL Task() calls in a SINGLE MESSAGE for parallel execution!**
+
+For each agent, create a Task() call with subagent_type="domain-plugin-builder:agents-builder" that includes:
+- Agent name and description
+- Instructions to load templates from build-assistant/templates/agents/
+- File path: $BASE_PATH/agents/{agent-name}.md
+- Frontmatter: name, description, model: inherit, color (from agent-color-decision.md)
+- **NO tools field** - agents inherit from parent
+- Available Tools & Resources section (MCP servers, skills, slash commands)
+- Progressive WebFetch for docs
+- Keep under 300 lines
+- Validate with validation script
+
+**Send ALL Task() calls together in ONE message - they will execute in parallel!**
+
+Only proceed to Phase 4 after all Task() calls complete.
+
+Phase 4: Validation and Registration
+
+**Validate all created agents:**
+
+For each agent:
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-agent.sh $BASE_PATH/agents/$AGENT_NAME.md}
+
+If validation fails, read errors and fix issues.
+
+**Note:** Agents don't need settings.json registration (only commands do).
+
+Phase 5: Git Commit and Push
+Goal: Save work immediately
+
+Actions:
+- Add all agent files to git: !{bash git add plugins/*/agents/*.md}
+- Commit with message:
+  !{bash git commit -m "$(cat <<'EOF'
+feat: Add agent(s) - AGENT_NAMES
+
+Complete agent structure with framework compliance.
+
+🤖 Generated with [Claude Code](https://claude.com/claude-code)
+
+Co-Authored-By: Claude <noreply@anthropic.com>
+EOF
+)"}
+- Push to GitHub: !{bash git push origin master}
+
+Phase 6: Sync to Airtable
+Goal: Sync ALL created agents to Airtable in bulk
+
+Actions:
+
+**Use bulk sync script for efficiency:**
+
+Bash: python ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/scripts/bulk-sync-airtable.py --plugin={plugin-name} --marketplace={marketplace-name} --type=agents
+
+This syncs ALL agents in parallel instead of one at a time.
+
+**DO NOT skip this phase!** Airtable sync is critical for marketplace integration.
+
+Phase 7: Self-Validation
+
+Run validation script to verify all work completed:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-agent.sh $BASE_PATH/agents/*.md}
+
+Mark all todos complete if validation passes.
+
+Phase 8: Summary
+Goal: Report results
+
+Actions:
+
+Display results:
+
+**Agents Created:** <count>
+**Plugin:** $PLUGIN_NAME
+**Mode:** $BASE_PATH (marketplace mode if "plugins/", standalone if ".")
+**Location:** $BASE_PATH/agents/
+
+**Files:**
+- $AGENT_1.md - $DESC_1 ✅
+- $AGENT_2.md - $DESC_2 ✅
+- etc.
+
+**Validation:** All passed ✅
+**Git Status:** Committed and pushed ✅
+**Airtable Sync:** Attempted ✅
+
+**Next Steps:**
+- Invoke agents via Task tool: Task(subagent_type="$PLUGIN_NAME:$AGENT_NAME")
+- Use in commands and skills
+- Test agent capabilities
--- a/commands/hooks-create.md
+++ b/commands/hooks-create.md
@@ -0,0 +1,244 @@
+---
+description: Create hook(s) following standardized structure - supports parallel creation for 3+ hooks
+argument-hint: <hook-name> <event-type> "<action>" [--plugin=name] [--marketplace] | <hook-1> <event-1> ... [--plugin=name] [--marketplace]
+---
+
+---
+🚨 **EXECUTION NOTICE FOR CLAUDE**
+
+When you invoke this command via `SlashCommand(/domain-plugin-builder:hooks-create ...)`, the system returns THESE INSTRUCTIONS below.
+
+**YOU are the executor. This is NOT an autonomous subprocess.**
+
+- ✅ The phases below are YOUR execution checklist
+- ✅ YOU must run each phase immediately using tools (Bash, Read, Write, Edit, TodoWrite)
+- ✅ Complete ALL phases before considering this command done
+- ❌ DON'T wait for "the command to complete" - YOU complete it by executing the phases
+- ❌ DON'T treat this as status output - it IS your instruction set
+
+**Immediately after SlashCommand returns, start executing Phase 0, then Phase 1, etc.**
+
+See `@CLAUDE.md` section "SlashCommand Execution - YOU Are The Executor" for detailed explanation.
+
+---
+
+## Security Requirements
+
+**CRITICAL:** All generated files must follow security rules:
+
+@~/.claude/plugins/marketplaces/dev-lifecycle-marketplace/docs/security/SECURITY-RULES.md
+
+**Key requirements:**
+- Never hardcode API keys or secrets
+- Use placeholders: `your_service_key_here`
+- Protect `.env` files with `.gitignore`
+- Create `.env.example` with placeholders only
+- Document key acquisition for users
+
+**Arguments**: $ARGUMENTS
+
+Goal: Create properly structured hooks. For 3+ hooks, creates them in parallel for faster execution.
+
+**CRITICAL EXECUTION INSTRUCTIONS:**
+- DO NOT wait for phases to run automatically
+- DO NOT just explain what the phases do
+- EXECUTE each phase immediately using the actual tools (Bash, Read, Write, Edit, TodoWrite, Task)
+- The `!{tool command}` syntax shows you WHAT to execute - use the real tool to DO IT
+- Complete all phases in order before finishing
+
+Phase 0: Create Todo List
+
+Create todo list for all phases below.
+
+Phase 1: Load Architectural Framework
+
+Actions:
+- Load component decision guidance:
+  !{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/docs/frameworks/claude/reference/component-decision-framework.md}
+- Load composition patterns:
+  !{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/docs/frameworks/claude/reference/dans-composition-pattern.md}
+- These provide critical understanding of:
+  - Hooks are for event-driven automation
+  - When to use hooks vs commands
+  - Hook boundaries and responsibilities
+  - How hooks integrate with plugin ecosystem
+  - Anti-patterns to avoid
+
+Phase 2: Parse Arguments and Determine Plugin & Mode
+
+Parse $ARGUMENTS to extract:
+- Hook names and event types
+- Plugin name (from --plugin=name or detect from pwd)
+- Marketplace mode (check for --marketplace flag)
+
+If plugin not specified:
+
+!{bash basename $(pwd)}
+
+Determine base path - check if already in a plugin directory:
+
+!{bash test -f .claude-plugin/plugin.json && echo "." || (echo "$ARGUMENTS" | grep -q "\-\-marketplace" && echo "plugins/$(basename $(pwd))" || echo ".")}
+
+Store as $BASE_PATH:
+- If .claude-plugin/plugin.json exists: BASE_PATH="." (already in plugin directory)
+- Else if --marketplace present: BASE_PATH="plugins/$PLUGIN_NAME"
+- Else: BASE_PATH="." (standalone plugin mode)
+
+All subsequent file operations use $BASE_PATH instead of hardcoded "plugins/$PLUGIN_NAME"
+
+Phase 3: Load Hooks Documentation
+
+WebFetch: https://docs.claude.com/en/docs/claude-code/hooks-guide
+
+This provides context on:
+- Available event types and when they trigger
+- Hook configuration structure
+- Script patterns and best practices
+
+Phase 4: Parse Arguments & Determine Mode
+
+Actions:
+
+Use bash to parse $ARGUMENTS and count how many hooks are being requested:
+
+!{bash echo "$ARGUMENTS" | grep -oE '<[^>]+>' | wc -l}
+
+Store the count. Then extract each hook specification:
+- If count = 1: Single hook mode - extract <hook-name>, <event-type>, and "<action>"
+- If count = 2: Two hooks mode - extract both sets
+- If count >= 3: Multiple hooks mode - extract all sets
+
+Execution modes:
+- 1 hook: Direct creation
+- 2 hooks: Sequential creation
+- 3+ hooks: Parallel creation (invoke multiple hooks-builder agents)
+
+Phase 5: Create Hooks
+
+**Mode 1: Single Hook (1 hook)**
+
+Task(description="Create hook", subagent_type="domain-plugin-builder:hooks-builder", prompt="You are the hooks-builder agent. Create a complete hook.
+
+Hook name: <name-from-args>
+Event type: <event-from-args>
+Action: <action-from-args>
+Plugin: <plugin-name>
+
+Create hook configuration and script:
+- Executable script in $BASE_PATH/scripts/
+- Hook entry in $BASE_PATH/hooks/hooks.json
+- Documentation in $BASE_PATH/docs/hooks.md
+- Use ${CLAUDE_PLUGIN_ROOT} for paths
+- Validate hook structure
+
+Deliverable: Complete hook with script, config, and documentation")
+
+**Mode 2: Sequential (2 hooks)**
+
+For each hook sequentially:
+Task(description="Create hook N", subagent_type="domain-plugin-builder:hooks-builder", prompt="<same structure>")
+
+Wait for completion, then create next hook.
+
+**Mode 3: Parallel (3+ hooks)**
+
+Create TODO list:
+
+TodoWrite with list of all hooks to create.
+
+Launch ALL hooks-builder agents IN PARALLEL (all at once):
+
+Task(description="Create hook 1", subagent_type="domain-plugin-builder:hooks-builder", prompt="You are the hooks-builder agent. Create a complete hook.
+
+Hook name: $HOOK_1
+Event type: $EVENT_1
+Action: $ACTION_1
+Plugin: $PLUGIN_NAME
+
+Create:
+- Script: $BASE_PATH/scripts/$HOOK_1.sh
+- Config: Update $BASE_PATH/hooks/hooks.json
+- Docs: Update $BASE_PATH/docs/hooks.md
+- Use ${CLAUDE_PLUGIN_ROOT} for all paths
+
+Deliverable: Complete hook")
+
+Task(description="Create hook 2", subagent_type="domain-plugin-builder:hooks-builder", prompt="Create hook: $HOOK_2 - Event: $EVENT_2 [same prompt structure as hook 1 above]")
+
+Task(description="Create hook 3", subagent_type="domain-plugin-builder:hooks-builder", prompt="Create hook: $HOOK_3 - Event: $EVENT_3 [same prompt structure as hook 1 above]")
+
+[Continue for all N hooks requested]
+
+Each Task() call happens in parallel. Parse $ARGUMENTS to determine how many Task() calls to make.
+
+Wait for ALL agents to complete before proceeding.
+
+Update TodoWrite as each completes.
+
+Phase 5.5: Git Commit and Push
+Goal: Save work immediately
+
+Actions:
+- Add all hook files to git:
+  !{bash git add $BASE_PATH/hooks/ $BASE_PATH/scripts/ $BASE_PATH/docs/hooks.md}
+- Commit with message:
+  !{bash git commit -m "$(cat <<'EOF'
+feat: Add hook(s) - HOOK_NAMES
+
+Complete hook structure with scripts, config, and documentation.
+
+🤖 Generated with [Claude Code](https://claude.com/claude-code)
+
+Co-Authored-By: Claude <noreply@anthropic.com>
+EOF
+)"}
+- Push to GitHub: !{bash git push origin master}
+
+Phase 5.6: Sync to Airtable
+Goal: Sync ALL created hooks to Airtable in bulk
+
+Actions:
+
+**Use bulk sync script for efficiency:**
+
+Bash: python ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/scripts/bulk-sync-airtable.py --plugin={plugin-name} --marketplace={marketplace-name} --type=hooks
+
+This syncs ALL hooks in parallel instead of one at a time.
+
+**Environment Requirement:**
+- Requires AIRTABLE_TOKEN environment variable
+- If not set, displays error message with instructions
+
+Phase 6: Self-Validation
+
+Run validation checks to verify all work completed:
+
+!{bash test -f $BASE_PATH/hooks/hooks.json && echo "✅ hooks.json exists" || echo "❌ hooks.json MISSING"}
+!{bash test -d $BASE_PATH/scripts && echo "✅ scripts/ exists" || echo "❌ scripts/ MISSING"}
+
+Mark all todos complete if validation passes.
+
+Phase 7: Summary
+Goal: Report results
+
+Actions:
+- Display results:
+
+**Hooks Created:** <count>
+**Plugin:** <plugin-name>
+**Mode:** $BASE_PATH (marketplace mode if "plugins/", standalone if ".")
+**Location:** $BASE_PATH/hooks/
+
+**Hooks:**
+- <hook-1-name> (Event: <event-1>) - <action-1> ✅
+- <hook-2-name> (Event: <event-2>) - <action-2> ✅
+- etc.
+
+**Validation:** All passed ✅
+**Git Status:** Committed and pushed ✅
+**Airtable Sync:** Attempted ✅
+
+**Next Steps:**
+- Test hooks by triggering events
+- Review scripts for correctness
+- Monitor hook execution in logs
--- a/commands/plugin-create.md
+++ b/commands/plugin-create.md
@@ -0,0 +1,190 @@
+---
+description: Create basic plugin directory structure and manifest
+argument-hint: <plugin-name> [--marketplace]
+---
+
+---
+🚨 **EXECUTION NOTICE FOR CLAUDE**
+
+When you invoke this command via `SlashCommand(/domain-plugin-builder:plugin-create ...)`, the system returns THESE INSTRUCTIONS below.
+
+**YOU are the executor. This is NOT an autonomous subprocess.**
+
+- ✅ The phases below are YOUR execution checklist
+- ✅ YOU must run each phase immediately using tools (Bash, Read, Write, Edit, TodoWrite)
+- ✅ Complete ALL phases before considering this command done
+- ❌ DON'T wait for "the command to complete" - YOU complete it by executing the phases
+- ❌ DON'T treat this as status output - it IS your instruction set
+
+**Immediately after SlashCommand returns, start executing Phase 0, then Phase 1, etc.**
+
+See `@CLAUDE.md` section "SlashCommand Execution - YOU Are The Executor" for detailed explanation.
+
+---
+
+## Security Requirements
+
+**CRITICAL:** All generated files must follow security rules:
+
+@~/.claude/plugins/marketplaces/dev-lifecycle-marketplace/docs/security/SECURITY-RULES.md
+
+**Key requirements:**
+- Never hardcode API keys or secrets
+- Use placeholders: `your_service_key_here`
+- Protect `.env` files with `.gitignore`
+- Create `.env.example` with placeholders only
+- Document key acquisition for users
+
+**Arguments**: $ARGUMENTS
+
+Goal: Create the basic directory structure and plugin.json manifest for a new Claude Code plugin.
+
+This is a simple command focused on creating the scaffold. Use /domain-plugin-builder:build-plugin for complete orchestration.
+
+**CRITICAL EXECUTION INSTRUCTIONS:**
+- DO NOT wait for phases to run automatically
+- DO NOT just explain what the phases do
+- EXECUTE each phase immediately using the actual tools (Bash, Read, Write, Edit, TodoWrite, AskUserQuestion)
+- The `!{tool command}` syntax shows you WHAT to execute - use the real tool to DO IT
+- Complete all phases in order before finishing
+
+Phase 0: Create Todo List using TodoWrite tool
+
+Create todo list for all phases below.
+
+Phase 0.5: Load Template Files
+
+Read the template files to understand what variables need to be replaced:
+
+!{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/marketplace.json.template}
+
+!{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/plugin.json.template}
+
+!{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/README.md.template}
+
+This shows exactly what template variables exist and need to be replaced in Phase 4.
+
+Phase 1: Parse Arguments and Setup Structure
+
+Parse $ARGUMENTS to extract plugin name:
+
+!{bash echo "$ARGUMENTS" | sed 's/--marketplace//g' | xargs}
+
+Store as $PLUGIN_NAME.
+
+**CRITICAL: Always create marketplace structure**
+
+Structure to create:
+```
+$PLUGIN_NAME/
+├── .claude-plugin/
+│   └── marketplace.json    ← Marketplace manifest
+└── plugins/
+    └── $PLUGIN_NAME/
+        ├── .claude-plugin/
+        │   └── plugin.json  ← Plugin manifest
+        ├── commands/
+        ├── agents/
+        ├── skills/
+        └── ...
+```
+
+This matches domain-plugin-builder's structure exactly.
+
+Phase 2: Gather Basic Info
+
+Use AskUserQuestion to get:
+- Plugin description (one sentence)
+- Plugin type (SDK, Framework, Custom)
+- Author name (default: "Plugin Developer")
+- Author email (default: "noreply@example.com")
+- License (default: "MIT")
+- Repository URL (default: "https://github.com/username/{{PLUGIN_NAME}}")
+- Homepage URL (default: "https://github.com/username/{{PLUGIN_NAME}}")
+
+Store these values for Phase 4.
+
+Phase 3: Create Marketplace and Plugin Directory Structure
+
+Create marketplace root structure:
+
+!{bash mkdir -p $PLUGIN_NAME/.claude-plugin}
+
+Create plugin subdirectory structure:
+
+!{bash mkdir -p $PLUGIN_NAME/plugins/$PLUGIN_NAME/.claude-plugin $PLUGIN_NAME/plugins/$PLUGIN_NAME/commands $PLUGIN_NAME/plugins/$PLUGIN_NAME/agents $PLUGIN_NAME/plugins/$PLUGIN_NAME/skills $PLUGIN_NAME/plugins/$PLUGIN_NAME/hooks $PLUGIN_NAME/plugins/$PLUGIN_NAME/scripts $PLUGIN_NAME/plugins/$PLUGIN_NAME/docs}
+
+Result:
+- Marketplace root: `$PLUGIN_NAME/.claude-plugin/`
+- Plugin: `$PLUGIN_NAME/plugins/$PLUGIN_NAME/`
+
+Phase 4: Create Marketplace and Plugin Files from Templates
+
+Set template variables:
+- PLUGIN_NAME=<from Phase 1>
+- DESCRIPTION=<from Phase 2>
+- AUTHOR_NAME=<from Phase 2>
+- AUTHOR_EMAIL=<from Phase 2>
+- LICENSE=<from Phase 2>
+- REPOSITORY_URL=<from Phase 2>
+- HOMEPAGE_URL=<from Phase 2>
+- VERSION="1.0.0"
+- KEYWORDS='["plugin", "claude-code"]'
+- DATE=$(date +%Y-%m-%d)
+
+**Create marketplace.json at root:**
+
+!{bash sed "s/{{PLUGIN_NAME}}/$PLUGIN_NAME/g; s/{{DESCRIPTION}}/$DESCRIPTION/g" ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/marketplace.json.template > $PLUGIN_NAME/.claude-plugin/marketplace.json}
+
+**Create plugin.json in plugins subdirectory:**
+
+Replace ALL template variables in plugin.json:
+
+!{bash sed "s/{{PLUGIN_NAME}}/$PLUGIN_NAME/g; s/{{VERSION}}/$VERSION/g; s/{{DESCRIPTION}}/$DESCRIPTION/g; s/{{AUTHOR_NAME}}/$AUTHOR_NAME/g; s/{{AUTHOR_EMAIL}}/$AUTHOR_EMAIL/g; s/{{HOMEPAGE_URL}}/$HOMEPAGE_URL/g; s/{{REPOSITORY_URL}}/$REPOSITORY_URL/g; s/{{LICENSE}}/$LICENSE/g; s/{{KEYWORDS}}/$KEYWORDS/g" ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/plugin.json.template > $PLUGIN_NAME/plugins/$PLUGIN_NAME/.claude-plugin/plugin.json}
+
+**Create plugin files:**
+
+!{bash sed "s/{{PLUGIN_NAME}}/$PLUGIN_NAME/g; s/{{DESCRIPTION}}/$DESCRIPTION/g" ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/README.md.template > $PLUGIN_NAME/plugins/$PLUGIN_NAME/README.md}
+
+!{bash sed "s/{{DATE}}/$DATE/g" ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/CHANGELOG.md.template > $PLUGIN_NAME/plugins/$PLUGIN_NAME/CHANGELOG.md}
+
+!{bash cp ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/hooks.json.template $PLUGIN_NAME/plugins/$PLUGIN_NAME/hooks/hooks.json}
+
+!{bash cp ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/.gitignore.template $PLUGIN_NAME/plugins/$PLUGIN_NAME/.gitignore}
+
+!{bash cp ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/mcp.json.template $PLUGIN_NAME/plugins/$PLUGIN_NAME/.mcp.json}
+
+!{bash cp ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/plugins/LICENSE.template $PLUGIN_NAME/plugins/$PLUGIN_NAME/LICENSE}
+
+Phase 5: Self-Validation
+
+Run validation script to verify all work completed:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-plugin.sh $BASE_PATH}
+
+Mark all todos complete if validation passes.
+
+Phase 6: Summary
+
+Display:
+- Plugin created: $PLUGIN_NAME ✅
+- Mode: $BASE_PATH (marketplace mode if "plugins/", standalone if ".")
+- Location: $BASE_PATH
+- Files created:
+  - .claude-plugin/plugin.json ✅
+  - hooks/hooks.json ✅
+  - .gitignore ✅
+  - .mcp.json ✅
+  - LICENSE ✅
+  - CHANGELOG.md ✅
+  - README.md ✅
+- Directory structure: Complete ✅
+- Validation: Passed ✅
+- Next steps:
+  - Build components in this order:
+    1. Agents first: /domain-plugin-builder:agents-create <agent-name> "description"
+    2. Commands second: /domain-plugin-builder:slash-commands-create <cmd-name> "description"
+    3. Skills third: /domain-plugin-builder:skills-create <skill-name> "description"
+    4. Hooks last: /domain-plugin-builder:hooks-create <hook-name> <event> "action"
+  - Why agents first? Commands often invoke agents, so agents must exist first
+  - Validate when done: /domain-plugin-builder:validate $PLUGIN_NAME
--- a/commands/skills-create.md
+++ b/commands/skills-create.md
@@ -0,0 +1,254 @@
+---
+description: Create new skill(s) using skills-builder agent - analyzes plugin structure or accepts direct specifications (supports parallel creation)
+argument-hint: [--analyze <plugin-name>] | [<skill-name> "<description>"] | [<skill-1> "<desc-1>" <skill-2> "<desc-2>" ...] [--marketplace]
+---
+
+---
+🚨 **EXECUTION NOTICE FOR CLAUDE**
+
+When you invoke this command via `SlashCommand(/domain-plugin-builder:skills-create ...)`, the system returns THESE INSTRUCTIONS below.
+
+**YOU are the executor. This is NOT an autonomous subprocess.**
+
+- ✅ The phases below are YOUR execution checklist
+- ✅ YOU must run each phase immediately using tools (Bash, Read, Write, Edit, TodoWrite)
+- ✅ Complete ALL phases before considering this command done
+- ❌ DON'T wait for "the command to complete" - YOU complete it by executing the phases
+- ❌ DON'T treat this as status output - it IS your instruction set
+
+**Immediately after SlashCommand returns, start executing Phase 0, then Phase 1, etc.**
+
+See `@CLAUDE.md` section "SlashCommand Execution - YOU Are The Executor" for detailed explanation.
+
+---
+
+## Security Requirements
+
+**CRITICAL:** All generated files must follow security rules:
+
+@~/.claude/plugins/marketplaces/dev-lifecycle-marketplace/docs/security/SECURITY-RULES.md
+
+**Key requirements:**
+- Never hardcode API keys or secrets
+- Use placeholders: `your_service_key_here`
+- Protect `.env` files with `.gitignore`
+- Create `.env.example` with placeholders only
+- Document key acquisition for users
+
+**Arguments**: $ARGUMENTS
+
+Goal: Create properly structured skill(s) by launching the skills-builder agent
+
+**CRITICAL EXECUTION INSTRUCTIONS:**
+- DO NOT wait for phases to run automatically
+- DO NOT just explain what the phases do
+- EXECUTE each phase immediately using the actual tools (Bash, Read, Write, Edit, TodoWrite, Task)
+- The `!{tool command}` syntax shows you WHAT to execute - use the real tool to DO IT
+- Complete all phases in order before finishing
+
+Phase 0: Create Todo List
+
+Create todo list for all phases below.
+
+Phase 1: Discovery & Architecture Decision Framework
+Goal: Load comprehensive component decision framework to understand WHEN to use skills vs commands vs agents vs hooks vs MCP
+
+Actions:
+- Load the complete component decision framework:
+  !{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/docs/frameworks/claude/reference/component-decision-framework.md}
+- This provides critical understanding of:
+  - 🚨 START WITH COMMANDS FIRST (not skills!)
+  - Commands are the primitive (closest to prompts)
+  - Skills are for MANAGING multiple related commands
+  - The "One-Off vs Management" test
+  - When NOT to create a skill
+  - Real composition hierarchy (skills use commands, not vice versa)
+  - Complete decision tree with real examples
+  - Anti-patterns to avoid
+- This architectural decision-making context will be passed to the skills-builder agent
+- Agent will understand when the requested functionality should be a COMMAND instead of a skill
+
+Phase 2: Parse Arguments & Determine Mode
+
+Actions:
+
+Parse $ARGUMENTS to extract:
+- Skill names and descriptions
+- Plugin name (from --plugin=name or detect from pwd)
+- Marketplace mode (check for --marketplace flag)
+
+If plugin not specified:
+
+!{bash basename $(pwd)}
+
+Determine base path - check if already in a plugin directory:
+
+!{bash test -f .claude-plugin/plugin.json && echo "." || (echo "$ARGUMENTS" | grep -q "\-\-marketplace" && echo "plugins/$(basename $(pwd))" || echo ".")}
+
+Store as $BASE_PATH:
+- If .claude-plugin/plugin.json exists: BASE_PATH="." (already in plugin directory)
+- Else if --marketplace present: BASE_PATH="plugins/$PLUGIN_NAME"
+- Else: BASE_PATH="." (standalone plugin mode)
+
+All subsequent file operations use $BASE_PATH instead of hardcoded "plugins/$PLUGIN_NAME"
+
+Use bash to parse $ARGUMENTS and count how many skills are being requested:
+
+!{bash echo "$ARGUMENTS" | grep -oE '<[^>]+>' | wc -l}
+
+Store the count. Then extract each skill specification:
+- If count = 0 and --analyze present: Set mode to "analyze", extract plugin name
+- If count = 1: Single skill mode - extract <skill-name> and "<description>"
+- If count >= 2: Multiple skills mode - extract all <skill-N> "<desc-N>" pairs
+
+Phase 3: Launch Skills Builder Agent(s)
+
+Actions:
+
+**For --analyze mode:**
+
+Task(description="Analyze plugin for needed skills", subagent_type="domain-plugin-builder:skills-builder", prompt="You are the skills-builder agent. Analyze the plugin structure at $BASE_PATH to determine what skills are needed.
+
+Architectural context from component-decision-framework.md:
+@~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/docs/frameworks/claude/component-decision-framework.md
+
+Tasks:
+1. Read detailed skills documentation via WebFetch:
+   - https://docs.claude.com/en/docs/agents-and-tools/agent-skills/quickstart
+   - https://docs.claude.com/en/docs/agents-and-tools/agent-skills/best-practices
+   - https://docs.claude.com/en/docs/claude-code/skills
+   - https://docs.claude.com/en/docs/claude-code/slash-commands#skills-vs-slash-commands
+   - https://github.com/anthropics/claude-cookbooks/tree/main/skills
+   - https://www.anthropic.com/engineering/equipping-agents-for-the-real-world-with-agent-skills
+2. Analyze plugin commands and agents
+3. Identify what reusable capabilities are needed
+4. Report recommended skills to create
+
+Plugin: $BASE_PATH
+Deliverable: List of recommended skills with descriptions")
+
+**Decision: 1-2 skills = build directly, 3+ skills = use Task() for parallel**
+
+**For 1-2 Skills:**
+
+Build directly - execute these steps immediately:
+
+1. Load decision framework:
+!{Read @component-decision-framework.md}
+
+2. Load skill template:
+!{Read @SKILL.md.template}
+
+3. For each skill:
+!{Bash mkdir -p $BASE_PATH/skills/$SKILL_NAME/{scripts,templates,examples}}
+
+!{Write $BASE_PATH/skills/$SKILL_NAME/SKILL.md}
+
+Create scripts, templates, and examples as needed.
+
+4. Validate:
+!{Bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-skill.sh $BASE_PATH/skills/$SKILL_NAME}
+
+No need for Task() overhead when building 1-2 skills
+
+**For 3+ Skills:**
+
+Launch multiple skills-builder agents IN PARALLEL (all at once) using multiple Task() calls in ONE response:
+
+Task(description="Create skill 1", subagent_type="domain-plugin-builder:skills-builder", prompt="Create skill: $SKILL_1 - $DESC_1 [same prompt structure as single skill above]")
+
+Task(description="Create skill 2", subagent_type="domain-plugin-builder:skills-builder", prompt="Create skill: $SKILL_2 - $DESC_2 [same prompt structure as single skill above]")
+
+Task(description="Create skill 3", subagent_type="domain-plugin-builder:skills-builder", prompt="Create skill: $SKILL_3 - $DESC_3 [same prompt structure as single skill above]")
+
+[Continue for all N skills requested]
+
+Each Task() call happens in parallel. Parse $ARGUMENTS to determine how many Task() calls to make.
+
+Wait for ALL agents to complete before proceeding to Phase 4.
+
+Phase 4: Validation
+
+**Validate all created skills:**
+
+For each skill:
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-skill.sh $BASE_PATH/skills/$SKILL_NAME}
+
+If validation fails, read errors and fix issues.
+
+Phase 5: Sync to Airtable
+
+**Use bulk sync script for efficiency:**
+
+Bash: python ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/scripts/bulk-sync-airtable.py --plugin={plugin-name} --marketplace={marketplace-name} --type=skills
+
+This syncs ALL skills in parallel instead of one at a time.
+
+**Environment Requirement:**
+- Requires AIRTABLE_TOKEN environment variable
+- If not set, displays error message with instructions
+
+Phase 6: Register Skills in Settings
+
+**CRITICAL: Skills must be registered to be usable!**
+
+Register all created skills in ~/.claude/settings.json:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/register-skills-in-settings.sh $PLUGIN_NAME}
+
+This registers entries like:
+- Skill($PLUGIN_NAME:$SKILL_1)
+- Skill($PLUGIN_NAME:$SKILL_2)
+- etc.
+
+Verify skills are accessible with Skill tool.
+
+Phase 7: Git Commit and Push
+
+Actions:
+- Add all created skill directories to git:
+  !{bash git add plugins/*/skills/*}
+- Commit with descriptive message:
+  !{bash git commit -m "$(cat <<'EOF'
+feat: Add skill(s) - SKILL_NAMES
+
+Complete skill structure with scripts, templates, and examples.
+
+🤖 Generated with [Claude Code](https://claude.com/claude-code)
+
+Co-Authored-By: Claude <noreply@anthropic.com>
+EOF
+)"}
+- Push to GitHub: !{bash git push origin master}
+
+Phase 8: Self-Validation
+
+Run validation script to verify all work completed:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-skill.sh $BASE_PATH/skills/*}
+
+Mark all todos complete if validation passes.
+
+Phase 9: Summary
+
+Display results:
+
+**Skills Created:** <count>
+**Plugin:** $PLUGIN_NAME
+**Mode:** $BASE_PATH (marketplace mode if "plugins/", standalone if ".")
+**Location:** $BASE_PATH/skills/
+
+**Files:**
+- $SKILL_1/ - $DESC_1 ✅
+- $SKILL_2/ - $DESC_2 ✅
+- etc.
+
+**Validation:** All passed ✅
+**Registration:** Complete ✅
+**Airtable Sync:** Attempted ✅
+**Git:** Committed and pushed ✅
+
+**Next Steps:**
+- Invoke skills: Skill($PLUGIN_NAME:$SKILL_NAME)
+- Test skill functionality
+- Update documentation
--- a/commands/slash-commands-create.md
+++ b/commands/slash-commands-create.md
@@ -0,0 +1,241 @@
+---
+description: Create slash command(s) following standardized structure - supports parallel creation for 3+ commands
+argument-hint: <command-name> "<description>" [--plugin=name] | <cmd-1> "<desc-1>" <cmd-2> "<desc-2>" ... [--plugin=name]
+---
+
+---
+🚨 **EXECUTION NOTICE FOR CLAUDE**
+
+When you invoke this command via `SlashCommand(/domain-plugin-builder:slash-commands-create ...)`, the system returns THESE INSTRUCTIONS below.
+
+**YOU are the executor. This is NOT an autonomous subprocess.**
+
+- ✅ The phases below are YOUR execution checklist
+- ✅ YOU must run each phase immediately using tools (Bash, Read, Write, Edit, TodoWrite)
+- ✅ Complete ALL phases before considering this command done
+- ❌ DON'T wait for "the command to complete" - YOU complete it by executing the phases
+- ❌ DON'T treat this as status output - it IS your instruction set
+
+**Immediately after SlashCommand returns, start executing Phase 0, then Phase 1, etc.**
+
+See `@CLAUDE.md` section "SlashCommand Execution - YOU Are The Executor" for detailed explanation.
+
+---
+
+## Security Requirements
+
+**CRITICAL:** All generated files must follow security rules:
+
+@~/.claude/plugins/marketplaces/dev-lifecycle-marketplace/docs/security/SECURITY-RULES.md
+
+**Key requirements:**
+- Never hardcode API keys or secrets
+- Use placeholders: `your_service_key_here`
+- Protect `.env` files with `.gitignore`
+- Create `.env.example` with placeholders only
+- Document key acquisition for users
+
+**Arguments**: $ARGUMENTS
+
+Goal: Create properly structured slash command(s). For 3+ commands, creates them in parallel for faster execution.
+
+**CRITICAL EXECUTION INSTRUCTIONS:**
+- DO NOT wait for phases to run automatically
+- DO NOT just explain what the phases do
+- EXECUTE each phase immediately using the actual tools (Bash, Read, Write, Edit, TodoWrite)
+- The `!{tool command}` syntax shows you WHAT to execute - use the real tool to DO IT
+- Complete all phases in order before finishing
+
+Phase 0: Create Todo List
+
+Create todo list for all phases below.
+
+Phase 1: Load Architectural Framework
+
+Actions:
+- Load component decision guidance:
+  !{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/docs/frameworks/claude/reference/component-decision-framework.md}
+- Load composition patterns:
+  !{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/docs/frameworks/claude/reference/dans-composition-pattern.md}
+- These provide critical understanding of:
+  - Commands are the primitive (start here!)
+  - When to use commands vs skills vs agents vs hooks
+  - Skills compose commands (not vice versa)
+  - Composition hierarchies and patterns
+  - Anti-patterns to avoid
+
+Phase 2: Parse Arguments and Determine Plugin
+
+Parse $ARGUMENTS to extract:
+- Command names and descriptions
+- Plugin name (from --plugin=name or detect from pwd)
+
+If plugin not specified:
+
+!{bash basename $(pwd)}
+
+Store plugin name for Phase 3.
+
+Phase 3: Load Templates
+
+Load command template for reference:
+
+!{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/commands/template-command-patterns.md}
+
+Phase 4: Parse Arguments & Determine Mode
+
+Actions:
+
+Use bash to parse $ARGUMENTS and count how many commands are being requested:
+
+!{bash echo "$ARGUMENTS" | grep -oE '<[^>]+>' | wc -l}
+
+Store the count. Then extract each command specification:
+- If count = 1: Single command mode - extract <command-name> and "<description>"
+- If count = 2: Two commands mode - extract both <cmd-N> "<desc-N>" pairs
+- If count >= 3: Multiple commands mode - extract all <cmd-N> "<desc-N>" pairs
+
+Phase 5: Create Commands
+
+Actions:
+
+**Decision: 1-2 commands = build directly, 3+ commands = use Task() for parallel**
+
+**For 1-2 Commands:**
+
+Build directly - execute these steps immediately:
+
+1. **List existing agents (to use correct names in Task() calls):**
+!{ls plugins/$PLUGIN_NAME/agents/*.md 2>/dev/null}
+
+2. Load template:
+!{Read ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/templates/commands/template-command-patterns.md}
+
+3. For each command, create the file:
+!{Write plugins/$PLUGIN_NAME/commands/$CMD_NAME.md}
+
+Include:
+- Frontmatter with description, argument-hint, allowed-tools
+- Use Goal → Actions → Phase pattern
+- Keep under 150 lines
+- **CRITICAL:** If command invokes agents via Task(), use ACTUAL agent names from step 1!
+  - Format: subagent_type="$PLUGIN_NAME:agent-name"
+  - Example: If agents/security-scanner.md exists, use subagent_type="my-plugin:security-scanner"
+  - NEVER use placeholder names - use REAL agent file names!
+
+4. Validate:
+!{Bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-command.sh plugins/$PLUGIN_NAME/commands/$CMD_NAME.md}
+
+No need for Task() overhead when building 1-2 commands
+
+**For 3+ Commands:**
+
+**CRITICAL: Send ALL Task() calls in a SINGLE MESSAGE for parallel execution!**
+
+Example for 3 commands - send all at once:
+
+```
+Task(description="Create command 1", subagent_type="domain-plugin-builder:slash-commands-builder", prompt="You are the slash-commands-builder agent. Create a complete slash command.
+
+Command name: $CMD_1_NAME
+Description: $CMD_1_DESC
+Plugin: $PLUGIN_NAME
+
+**CRITICAL: Use ACTUAL Agent Names**
+BEFORE creating the command, list existing agents:
+!{ls plugins/$PLUGIN_NAME/agents/*.md 2>/dev/null}
+
+If the command needs to invoke an agent via Task(), use the ACTUAL agent names from the list above!
+- Format: subagent_type=\"$PLUGIN_NAME:agent-name\"
+- Example: If agents/security-scanner.md exists, use subagent_type=\"$PLUGIN_NAME:security-scanner\"
+- NEVER use placeholder names like \"agent1\" or \"scanner\" - use the REAL agent file names!
+
+Load template: plugins/domain-plugin-builder/skills/build-assistant/templates/commands/template-command-patterns.md
+
+Create command file at: plugins/$PLUGIN_NAME/commands/$CMD_1_NAME.md
+
+Follow framework structure:
+- Frontmatter with description, argument-hint, allowed-tools
+- Goal → Actions → Phase pattern
+- Keep under 150 lines
+- Use $ARGUMENTS for all argument references
+- Use ACTUAL agent names in Task() calls (subagent_type field)
+- Validate with validation script
+
+Deliverable: Complete validated command file with correct agent references")
+
+Task(description="Create command 2", subagent_type="domain-plugin-builder:slash-commands-builder", prompt="[Same structure with $CMD_2_NAME, $CMD_2_DESC]")
+
+Task(description="Create command 3", subagent_type="domain-plugin-builder:slash-commands-builder", prompt="[Same structure with $CMD_3_NAME, $CMD_3_DESC]")
+
+[Continue for all N commands from $ARGUMENTS]
+```
+
+**DO NOT wait between Task() calls - send them ALL together in one response!**
+
+The agents will run in parallel automatically. Only proceed to Phase 6 after all Task() calls complete.
+
+Phase 6: Validation
+
+For each created command:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-command.sh plugins/$PLUGIN_NAME/commands/$CMD_NAME.md}
+
+If validation fails, read errors and fix issues.
+
+Phase 7: Sync to Airtable
+
+**Use bulk sync script for efficiency:**
+
+Bash: python ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/scripts/bulk-sync-airtable.py --plugin={plugin-name} --marketplace={marketplace-name} --type=commands
+
+This syncs ALL commands in parallel instead of one at a time.
+
+**Environment Requirement:**
+- Requires AIRTABLE_TOKEN environment variable
+- If not set, displays error message with instructions
+
+Phase 8: Register in Settings
+
+**CRITICAL: Commands must be registered to be usable!**
+
+Run registration script to add all created commands to ~/.claude/settings.json:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/register-commands-in-settings.sh $PLUGIN_NAME}
+
+This registers entries like:
+- SlashCommand(/$PLUGIN_NAME:$CMD_1)
+- SlashCommand(/$PLUGIN_NAME:$CMD_2)
+- etc.
+
+Verify registration by checking settings.json contains the new commands.
+
+Phase 9: Self-Validation
+
+Run validation script to verify all work completed:
+
+!{bash ~/.claude/plugins/marketplaces/domain-plugin-builder/plugins/domain-plugin-builder/skills/build-assistant/scripts/validate-command.sh plugins/$PLUGIN_NAME/commands/*.md}
+
+Mark all todos complete if validation passes.
+
+Phase 10: Summary
+
+Display results:
+
+**Commands Created:** <count>
+**Plugin:** $PLUGIN_NAME
+**Location:** plugins/$PLUGIN_NAME/commands/
+
+**Files:**
+- $CMD_1.md - $DESC_1 ✅
+- $CMD_2.md - $DESC_2 ✅
+- etc.
+
+**Validation:** All passed ✅
+**Registration:** Complete ✅
+**Airtable Sync:** Attempted ✅
+
+**Next Steps:**
+- Test commands: /$PLUGIN_NAME:$CMD_NAME
+- Add to build workflow
+- Commit to git
--- a/commands/validate.md
+++ b/commands/validate.md
@@ -0,0 +1,114 @@
+---
+description: Validate plugin structure and compliance using plugin-validator agent
+argument-hint: <plugin-name>
+---
+
+---
+🚨 **EXECUTION NOTICE FOR CLAUDE**
+
+When you invoke this command via `SlashCommand(/domain-plugin-builder:validate ...)`, the system returns THESE INSTRUCTIONS below.
+
+**YOU are the executor. This is NOT an autonomous subprocess.**
+
+- ✅ The phases below are YOUR execution checklist
+- ✅ YOU must run each phase immediately using tools (Bash, Read, Write, Edit, TodoWrite)
+- ✅ Complete ALL phases before considering this command done
+- ❌ DON'T wait for "the command to complete" - YOU complete it by executing the phases
+- ❌ DON'T treat this as status output - it IS your instruction set
+
+**Immediately after SlashCommand returns, start executing Phase 0, then Phase 1, etc.**
+
+See `@CLAUDE.md` section "SlashCommand Execution - YOU Are The Executor" for detailed explanation.
+
+---
+
+## Security Requirements
+
+**CRITICAL:** All generated files must follow security rules:
+
+@~/.claude/plugins/marketplaces/dev-lifecycle-marketplace/docs/security/SECURITY-RULES.md
+
+**Key requirements:**
+- Never hardcode API keys or secrets
+- Use placeholders: `your_service_key_here`
+- Protect `.env` files with `.gitignore`
+- Create `.env.example` with placeholders only
+- Document key acquisition for users
+
+**Arguments**: $ARGUMENTS
+
+Goal: Validate that a Claude Code plugin is properly structured, follows framework conventions, and is ready for deployment.
+
+**CRITICAL EXECUTION INSTRUCTIONS:**
+- DO NOT wait for phases to run automatically
+- DO NOT just explain what the phases do
+- EXECUTE each phase immediately using the actual tools (Bash, TodoWrite, Task)
+- The `!{tool command}` syntax shows you WHAT to execute - use the real tool to DO IT
+- Complete all phases in order before finishing
+
+Core Principles:
+- Invoke plugin-validator agent for comprehensive validation
+- Agent runs all validation scripts internally
+- Report validation results clearly
+
+Phase 0: Create Todo List
+
+Create todo list for all phases below.
+
+Phase 1: Verify Plugin Exists and Determine Mode
+
+Check if we're in a plugin directory or need to look in plugins/:
+
+!{bash test -f .claude-plugin/plugin.json && echo "standalone" || (test -d "plugins/$ARGUMENTS" && echo "marketplace" || echo "not-found")}
+
+Store mode:
+- If "standalone": PLUGIN_PATH="." (current directory is the plugin)
+- If "marketplace": PLUGIN_PATH="plugins/$ARGUMENTS"
+- If "not-found": Display error and exit
+
+If not found, display error:
+- "Plugin not found. Either:"
+- "  1. Run from inside plugin directory (with .claude-plugin/plugin.json)"
+- "  2. Run from marketplace root with plugin name as argument"
+
+Phase 2: Invoke Plugin Validator Agent
+
+Launch the plugin-validator agent to perform comprehensive validation:
+
+Task(description="Validate plugin structure and compliance", subagent_type="domain-plugin-builder:plugin-validator", prompt="You are the plugin-validator agent. Perform comprehensive validation on $PLUGIN_PATH.
+
+Run all validation scripts:
+- validate-plugin.sh for structure
+- validate-command.sh for each command
+- validate-agent.sh for each agent
+- validate-plugin-manifest.sh for plugin.json
+
+Verify:
+- Directory structure compliance
+- All agents follow framework conventions
+- All commands follow pattern templates
+- All skills have required directories
+- Tool formatting (no JSON arrays, no wildcards)
+- Line count limits (agents <300, commands <150)
+- Plugin manifest correctness
+- Documentation completeness
+
+Deliverable: Complete validation report with pass/fail status, critical issues, warnings, and recommendations.")
+
+Phase 3: Summary
+
+Display validation completion message:
+
+**Validation Complete for:** $ARGUMENTS
+
+See detailed report above from plugin-validator agent.
+
+If validation passed, next steps:
+- Plugin is ready for use
+- Commit changes if modifications were made
+- Push to GitHub: !{bash git push origin master}
+
+If validation failed:
+- Review critical issues in report
+- Fix issues identified
+- Re-run validation: /domain-plugin-builder:validate $ARGUMENTS