Compétences llm-application-dev-ai-assistant
📩

llm-application-dev-ai-assistant

Sûr

Build AI Assistants with Claude

Developers struggle to create production-ready AI assistants with proper conversation flow and NLP integration. This skill provides comprehensive patterns, code examples, and best practices for building intelligent conversational interfaces using Claude, Codex, and Claude Code.

Prend en charge: Claude Codex Code(CC)
đŸ„‰ 74 Bronze
1

Télécharger le ZIP du skill

2

Importer dans Claude

Allez dans ParamĂštres → CapacitĂ©s → Skills → Importer un skill

3

Activez et commencez Ă  utiliser

Tester

Utilisation de "llm-application-dev-ai-assistant". Design a conversation flow for a restaurant reservation assistant

RĂ©sultat attendu:

  • Greeting: 'Welcome! How can I help you today?'
  • Intent Detection: reservation_request
  • Slots needed: date, time, party_size, restaurant_name
  • Confirmation: 'Booking table for 4 at Italian Palace on March 15th at 7pm. Confirm?'
  • On confirm: Execute reservation function, provide confirmation number

Utilisation de "llm-application-dev-ai-assistant". Create an intent classifier for an e-commerce assistant

RĂ©sultat attendu:

  • Intent categories: product_search, price_inquiry, order_status, return_request, complaint
  • Confidence threshold: 0.75
  • Fallback: 'I did not understand. Could you rephrase?'
  • Context retention: Last 5 conversation turns

Audit de sécurité

Sûr
v1 ‱ 2/25/2026

This skill provides documentation and code examples for building AI assistants. All 59 static analyzer findings are false positives: (1) Backtick patterns are markdown code fences, not shell execution; (2) Method names like _design_dialog_manager incorrectly flagged as weak crypto; (3) 0.0.0.0:8080 is standard Docker bind address, not a vulnerability. No actual security risks detected.

2
Fichiers analysés
1,275
Lignes analysées
1
résultats
1
Total des audits
ProblĂšmes Ă  risque faible (1)
resources/implementation-playbook.md:22-1222SKILL.md:33-37
Static Analyzer False Positives
The static analyzer incorrectly flagged documentation patterns: (1) Markdown code fences using backticks were flagged as shell execution; (2) Method names with 'design' prefix were flagged as weak crypto; (3) Standard Docker bind address was flagged as hardcoded IP. These are all legitimate documentation patterns with no security risk.
Audité par: claude

Score de qualité

38
Architecture
100
Maintenabilité
87
Contenu
50
Communauté
100
Sécurité
91
Conformité aux spécifications

Ce que vous pouvez construire

Building Customer Support Chatbots

Create intelligent customer support assistants that understand user intent, maintain conversation context, and route queries to appropriate handlers.

Developing Internal productivity Assistants

Build AI assistants that help employees with tasks like scheduling, document retrieval, and process automation within an organization.

Creating Educational Tutoring Systems

Develop adaptive learning assistants that understand student queries, track progress, and provide personalized explanations.

Essayez ces prompts

Basic Assistant Architecture 
Design a basic AI assistant architecture for a customer support chatbot. Include components for intent recognition, dialog management, and response generation.
Conversation Flow Design 
Create a multi-turn conversation flow for handling order status inquiries. Include states for greeting, verification, status lookup, and resolution.
NLP Pipeline Implementation 
Implement an NLP pipeline that processes user messages through tokenization, intent classification, entity extraction, and sentiment analysis.
Function Calling Integration 
Design a function calling interface that allows an LLM to invoke external tools like database queries, API calls, and calendar operations.

Bonnes pratiques

  • Always implement proper error handling and graceful degradation when AI services are unavailable
  • Use context windows effectively by summarizing older conversation history
  • Test conversation flows with diverse user inputs including edge cases and unexpected queries

Éviter

  • Do not assume users will always provide complete information in a single message
  • Avoid blocking waiting for external API responses - use async patterns
  • Never hardcode sensitive credentials in assistant configuration

Foire aux questions

What programming languages are supported?
The code examples use Python, but the architectural patterns apply to any language with async capabilities.
Do I need machine learning experience?
Basic understanding of NLP concepts helps, but the skill provides implementation guidance for both rule-based and ML approaches.
Can this integrate with existing chatbot platforms?
Yes, the patterns can be adapted to integrate with platforms like Dialogflow, Rasa, or custom implementations.
How does context management work?
Context is maintained through a ConversationContext object that tracks user profile, conversation history, and session state.
What testing approaches are recommended?
Use unit tests for individual components, integration tests for conversation flows, and user simulation for end-to-end testing.
Can I use Claude Code with this skill?
Yes, this skill supports Claude, Codex, and Claude Code. The function calling patterns are designed for LLM integration.

Détails du développeur

Auteur

sickn33

Licence

MIT

DĂ©pĂŽt

https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/llm-application-dev-ai-assistant

RĂ©f

main

Structure de fichiers

📁 resources/

📄 implementation-playbook.md

📄 SKILL.md