smithery.ai

Agent Deployment

Production deployment workflow for agentic systems. Directs to RAG for implementation.

Installation

$ npx skills add https://smithery.ai

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery.ai · top by installs.

npx skills add https://smithery.ai

Browse all from smithery.ai

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 3,587 B
  • docs SUMMARY.md 110 B

History

  1. First recorded snapshot · 1 installs

SKILL.md

Agent Deployment Workflow

Deployment Decision Framework

Framework Primary Option Alternative RAG Query
ADK Agent Engine (Vertex AI) Cloud Run, GKE "ADK deployment agent engine"
OpenAI Any Python hosting Serverless, Docker "openai agents deployment"
LangChain LangServe, Cloud Run Docker, K8s "langchain langserve deployment"
LangGraph LangGraph Platform Cloud Run "langgraph platform deployment"
CrewAI CrewAI Enterprise Docker "crewai deployment production"
Anthropic Any Python hosting Docker, Serverless "anthropic agent deployment"

Pre-Deployment Checklist

Code Readiness

  • All tests passing
  • Error handling complete
  • Logging configured
  • Input validation in place
  • Output guardrails active

Configuration

  • Environment variables documented
  • Secrets in secret manager (NOT in .env or code)
  • Rate limiting configured
  • Token limits set
  • Timeout values appropriate

Security

  • API key rotation plan
  • Audit logging enabled
  • PII handling documented
  • Input sanitization active
  • Output filtering configured

Monitoring

  • Health check endpoint
  • Metrics collection
  • Alerting rules defined
  • Log aggregation setup

Deployment Workflow

Step 1: Environment Configuration

RAG Query: mcpagentic-ragsearch("[framework] environment configuration", mode="explain")

Production differs from development:

  • LOG_LEVEL: INFO (not DEBUG)
  • TRACE_ENABLED: false (or sampling)
  • Secrets: Use secret manager, not .env

Step 2: Containerization (if applicable)

RAG Query: mcpagentic-ragsearch("[framework] dockerfile", mode="build")

Step 3: Platform Deployment

RAG Query: mcpagentic-ragsearch("[framework] [platform] deployment", mode="explain")

Step 4: Monitoring Setup

RAG Query: mcpagentic-ragsearch("[framework] monitoring observability", mode="explain")

Key Production Metrics

Metric Alert Threshold Why It Matters
Latency p95 > 5s User experience
Error rate > 1% Reliability
Token usage Spike > 200% Cost control
Tool failures > 5% Agent effectiveness
Routing accuracy < 90% Multi-agent health

Security Considerations

Input Validation

  • Sanitize user input before passing to agent
  • Limit input length
  • Filter known attack patterns

RAG Query: mcpagentic-ragsearch("agent input validation security", mode="explain")

Output Guardrails

  • Filter sensitive information
  • Prevent prompt leakage
  • Validate tool outputs

RAG Query: mcpagentic-ragsearch("agent guardrails output filtering", mode="explain")

Secret Management

  • Never hardcode API keys
  • Use platform secret managers (GCP Secret Manager, AWS Secrets Manager, etc.)
  • Rotate keys regularly

RAG Query: mcpagentic-ragsearch("[framework] secret management", mode="explain")

Scaling Considerations

Concern Solution RAG Query
Cold starts Keep warm instances "[framework] cold start"
Concurrent requests Queue + workers "[framework] scaling"
Token limits Request batching "[framework] rate limiting"
State persistence External store "[framework] state persistence"