arch-advisor

Use this skill when designing the internal structure of a single agent, when discussing perception layers, decision layers, memory management, reflection loops, state management, or when someone says 'how should the agent be structured internally?', 'how do we handle memory?', 'should state be mutable or immutable?', 'we need the agent to self-improve its output', 'how do we chunk documents?', 'the agent needs to remember past interactions'. Also trigger for topics like LTM, STM, chunking strategy, critic-reviser, context window management.

Use this skill when documenting architectural decisions, creating ADRs, drawing C4 diagrams, writing RFCs, or presenting architecture to stakeholders. Trigger when someone says 'we need to document this decision', 'write an ADR', 'create a C4 diagram', 'how do we explain this to management?', 'we need an RFC', 'how do we communicate trade-offs?', 'design review', 'architecture presentation'. Also trigger for: Architecture Decision Records, C4 Model, trade-off analysis framework, design review, technical brief.

Use this skill when choosing storage systems, designing the data layer, deciding between databases, or designing agent memory. Trigger when someone says 'what database should we use?', 'where do we store agent state?', 'how do we persist memory between sessions?', 'we need vector storage for embeddings', 'how do we structure session data?', 'the agent needs to remember past conversations', 'we need to store user history', 'GDPR compliance for stored data'. Also trigger for: PostgreSQL, Redis, vector DB, Chroma, Pinecone, polyglot persistence, CQRS, Event Sourcing.

Use this skill when designing how agents expose or consume tools and APIs, when discussing MCP servers, A2A communication, tool registries, or when someone says 'agents need to call external APIs', 'how should agents communicate with each other?', 'we need to expose tools to the LLM', 'designing agent-to-agent messaging', 'tool governance and access control', 'idempotent tool calls', 'MCP server implementation'. Also trigger for: Model Context Protocol, A2A patterns, tool design, RBAC for tools, distributed tracing between agents.

Use this skill when integrating AI agents with existing enterprise systems, legacy APIs, ERPs, CRMs, or when migrating from legacy to agent-based systems. Trigger when someone says 'we have existing systems the agent must use', 'we need to integrate with our ERP/CRM', 'legacy system has a different data model', 'how do we migrate gradually?', 'our legacy system is unstable', 'we need to call a SOAP service', 'agents and legacy systems must coexist'. Also trigger for: Anti-Corruption Layer, Saga pattern, Strangler Fig migration, REST adapter, message queue integration.

Use this skill when choosing between LangChain, LangGraph, CrewAI, Semantic Kernel, or custom implementation, when someone asks 'which framework should we use?', 'should we use LangGraph or LangChain?', 'is CrewAI a good fit?', 'should we build custom?', 'we need a framework for our agents', 'comparing LLM frameworks', or when evaluating framework trade-offs for agent orchestration. Also trigger for topics: LCEL, StateGraph, Crew, Semantic Kernel skills, framework migration.

Use this skill when choosing which LLM to use, comparing providers, designing routing between models, handling LLM fallbacks, or optimizing LLM costs. Trigger when someone says 'which model should we use?', 'we need to choose between GPT-4 and Claude', 'how do we route to cheaper models for simple tasks?', 'we need fallback if OpenAI goes down', 'LLM costs are too high', 'we need multi-provider strategy', 'circuit breaker for LLM providers'. Also trigger for: model cascading, cost optimization, caching LLM responses, rate limiting.

Use this skill when designing systems with multiple agents, when choosing between centralized and decentralized coordination, when discussing pipelines, DAGs, agent communication, event-driven architectures, or when someone says 'we need multiple agents working together', 'how should agents coordinate?', 'should agents communicate directly or through a central controller?', 'we need agents to work in parallel', 'how do we handle agent conflicts?', 'we need a workflow with multiple steps and agents'. Also trigger for topics: orchestrator, choreography, Saga pattern, bounded contexts, agent contracts.

Use this skill when discussing monitoring, observability, SLOs, alerting, dashboards, or production reliability for AI systems. Trigger when someone says 'how do we monitor the agent?', 'we need alerts', 'how do we know if quality is degrading?', 'SLO for the system', 'the agent is slow and we don't know why', 'cost is unpredictable', 'we need dashboards', 'distributed tracing for agents'. Also trigger for: metrics, logs, traces, error budget, latency P95, hallucination monitoring, cost tracking.

Use this skill when designing customer-facing systems across multiple channels, when discussing chatbots, omnichannel, WhatsApp integration, web chat, email automation, or when someone says 'we need to serve customers on multiple channels', 'WhatsApp + web + email support', 'agent must adapt to each channel', 'we need session continuity across channels', 'intent classification for customer messages', 'escalation to human agent'. Also trigger for: channel adapters, intent routing, session management, response formatting by channel, human handoff.

Use this skill during Phase 3.5 Pattern Deepening to provide implementation-level guidance for the specific patterns present in the chosen architecture. Contains twelve pattern blocks: Hybrid Decision Engine, Planner-Executor-Critic, Voting + Arbiter, Saga with Compensation, Human-in-the-Loop with Checkpointing, Complexity-based LLM Routing, LLM Response Caching, Bulkhead, Anti-Corruption Layer, Strangler Fig, Batch Processing, Feedback Loop with Regression Detection.

Use this skill when designing RAG (Retrieval-Augmented Generation) pipelines, choosing chunking strategies, retrieval methods, embedding models, or reranking approaches. Trigger when someone says 'we need to search our documents', 'the agent needs to answer questions from our knowledge base', 'how should we chunk and index?', 'we need semantic search', 'the LLM is hallucinating facts from our docs', 'we need to cite sources', 'RAG is returning irrelevant results'. Also trigger for: vector databases, embeddings, hybrid search, multi-query retrieval.

Use this skill when discussing security, compliance, risk management, governance, or responsible AI for agent systems. Trigger when someone says 'we need LGPD compliance', 'how do we handle prompt injection?', 'what are the risks of using LLMs?', 'we need an audit trail', 'how do we govern AI models?', 'we process sensitive data', 'the agent can take real-world actions — what are the risks?', 'we need to explain agent decisions'. Also trigger for: TRiSM, AI Act, GDPR, hallucination risk, cost overrun, data leakage, bias detection.

Use this skill when designing testing strategies for AI agent systems, setting up quality evaluations, defining quality gates, or when someone says 'how do we test agents?', 'we need evals for LLM output', 'how to test non-deterministic systems?', 'quality gate for deployment', 'LLM-as-judge', 'regression testing for agents', 'how do we know if the model got worse?'. Also trigger for: testing pyramid, mock LLM, eval framework, hallucination detection, CI/CD for AI.

Use this skill when someone asks whether to build an agent or a traditional service, when to apply AI vs. deterministic logic, whether a use case justifies LLMs, or when a user says things like 'should we use AI here?', 'is this a good case for an agent?', 'we're deciding between an LLM and a rule-based system', 'does this need generative AI?', 'we need to classify tickets / interpret requests / generate responses'. Also trigger when discussing automation, chatbots, decision engines, or intelligent routing.