Load agent definitions on-demand to reduce context usage. Only loads full agent when needed.
/plugin marketplace add nguyenthienthanh/aura-frog/plugin install aura-frog@aurafrogThis skill is limited to using the following tools:
Priority: HIGH - Load agents on-demand Version: 1.0.0
Reduce token usage by:
agent_index[24]{id,category,specialty,keywords}:
mobile-react-native,dev,React Native/Expo mobile,react-native/expo/RN/mobile/ios/android
mobile-flutter,dev,Flutter/Dart mobile,flutter/dart/bloc/mobile
web-angular,dev,Angular frontend,angular/ngrx/rxjs/typescript
web-vuejs,dev,Vue.js frontend,vue/vuejs/pinia/nuxt/composition
web-reactjs,dev,React frontend,react/reactjs/jsx/hooks/redux
web-nextjs,dev,Next.js fullstack,next/nextjs/ssr/ssg/app-router
backend-nodejs,dev,Node.js backend,nodejs/express/nestjs/fastify/api
backend-python,dev,Python backend,python/django/fastapi/flask/api
backend-go,dev,Go backend,go/golang/gin/fiber/api
backend-laravel,dev,Laravel/PHP backend,laravel/php/eloquent/artisan
database-specialist,dev,Database/SQL,database/schema/query/migration/sql/postgres/mysql
security-expert,quality,Security auditing,security/vulnerability/audit/owasp/penetration
qa-automation,quality,Testing/QA,test/testing/coverage/qa/jest/cypress
ui-designer,quality,UI/UX design,design/ui/ux/layout/figma/style/component
devops-cicd,ops,DevOps/CI-CD,deploy/docker/kubernetes/ci-cd/pipeline/terraform
jira-operations,ops,JIRA integration,jira/ticket/issue/sprint/backlog
confluence-operations,ops,Confluence docs,confluence/wiki/documentation/page
slack-operations,ops,Slack notifications,slack/notification/channel/message
voice-operations,ops,Voice/TTS,voice/speak/tts/audio/notification
smart-agent-detector,infra,Agent detection,detect/agent/select/route
pm-operations-orchestrator,infra,PM/orchestration,pm/project/orchestrate/manage
project-detector,infra,Project detection,project/detect/identify/config
project-config-loader,infra,Config loading,config/load/settings/yaml
project-context-manager,infra,Context management,context/manage/state/session
1. Load this index file (agent_index)
2. DO NOT load individual agent files
3. Use index for agent detection scoring
1. Agent scores ≥80 (PRIMARY) → Load full definition
2. Agent scores 50-79 (SECONDARY) → Load summary only
3. Agent scores <50 (OPTIONAL) → Don't load
agents/[agent-id].md
# Load full agent definition
cat agents/mobile-react-native.md
agent_summary{id,role,focus}:
mobile-react-native,Senior React Native Developer,Expo/RN mobile apps with TypeScript
Step 1: Score agents using agent_index keywords
Step 2: Identify PRIMARY agent(s) with score ≥80
Step 3: Load ONLY PRIMARY agent full definitions
Step 4: For SECONDARY agents, use summary from index
comparison[4]{scenario,without_lazy,with_lazy,savings}:
Initial load (24 agents),~48000,~1200,97.5%
Single agent task,~48000,~2700,94.4%
Dual agent task,~48000,~4200,91.3%
Full stack (3 agents),~48000,~5700,88.1%
Loaded agents are cached in conversation context.
If agent already loaded, skip re-loading.
Track loaded agents: loaded_agents[]: mobile-react-native,qa-automation
User: "reload agent mobile-react-native"
→ Clear cache for agent
→ Re-read full definition
User: "Create a React Native screen for login"
1. Agent Detector scores all agents using keywords from agent_index
- mobile-react-native: +60 (react-native) +20 (context) = 80 ✅ PRIMARY
- ui-designer: +35 (screen/login implies UI) → OPTIONAL
2. Lazy Loader activates:
- Load: agents/mobile-react-native.md (~1500 tokens)
- Skip: ui-designer (score < 50)
3. Context loaded:
- Agent index: ~1200 tokens
- mobile-react-native full: ~1500 tokens
- Total: ~2700 tokens (vs ~48000 without lazy loading)
categories[4]{name,count,when_to_load}:
dev,11,When code/implementation requested
quality,3,When review/test/design requested
ops,5,When deploy/integrate/notify requested
infra,5,Usually auto-loaded by system
Note: This skill is automatically used by agent-detector for optimized loading.
This skill should be used when the user asks to "create an agent", "add an agent", "write a subagent", "agent frontmatter", "when to use description", "agent examples", "agent tools", "agent colors", "autonomous agent", or needs guidance on agent structure, system prompts, triggering conditions, or agent development best practices for Claude Code plugins.
This skill should be used when the user asks to "create a slash command", "add a command", "write a custom command", "define command arguments", "use command frontmatter", "organize commands", "create command with file references", "interactive command", "use AskUserQuestion in command", or needs guidance on slash command structure, YAML frontmatter fields, dynamic arguments, bash execution in commands, user interaction patterns, or command development best practices for Claude Code.
This skill should be used when the user asks to "create a hook", "add a PreToolUse/PostToolUse/Stop hook", "validate tool use", "implement prompt-based hooks", "use ${CLAUDE_PLUGIN_ROOT}", "set up event-driven automation", "block dangerous commands", or mentions hook events (PreToolUse, PostToolUse, Stop, SubagentStop, SessionStart, SessionEnd, UserPromptSubmit, PreCompact, Notification). Provides comprehensive guidance for creating and implementing Claude Code plugin hooks with focus on advanced prompt-based hooks API.