01
AI Agent
Architecture, tool use, memory, RAG, evaluation, permissions and runtime governance.
AI Engineering Knowledge Base
From architecture choices to production troubleshooting, this hub organizes real AI Agent, MCP, LangGraph, Workflow and on-device AI problems. Each topic aims to provide reproducible methods, constraints and a concrete next step.
⌕Search AI engineering problems, errors, frameworks, tools or production practice…Topics
02
Agent Security Infrastructure
Identity, authorization, approvals, sandboxing, MCP security, ZDR, audit and incident response for production agents.
03
MCP
Servers, transports, OAuth, permissions, security and protocol troubleshooting.
04
LangGraph
State, checkpoints, HITL, failure recovery, concurrency and observability.
05
Workflow Automation
n8n, self-hosting, webhooks, queues, retries, idempotency and cost governance.
06
AI Tools Lab
Model and AI-tool tests based on real projects and explicit tasks.
07
Local AI
iPhone, Android, Core ML, ONNX, llama.cpp, model size, heat, performance and privacy.
Guides & Practice
AI Agent Full Guide: From Architecture to Production Governance
Build the system boundary across agents, workflows, memory, RAG, tool use and production governance.
Read →
MCP vs Function Calling: Real Project Selection
Compare protocol boundaries, permissions, tool integration and client reuse.
Read →
LangGraph State Isolation in Practice
Design thread_id, session_id and user_id boundaries and persistence.
Read →
Self-hosted n8n AI Workflows
Docker, VPS, NAS, self-hosting and production operations.
Read →
AI Agent Tool Authorization Policy Gate
Authorization, approval, rejection and auditing before high-risk tool calls.
Read →
GPT-5.6 Real Project Test
Test whether model improvements materially change real development workflows.
Read →
Latest
What Is Context Engineering? Reducing AI Agent Cost with Retrieval, Tool Search and Memory
Context engineering goes beyond prompt engineering. This guide combines Microsoft, Anthropic and Google sources with XBSTACK tests on retrieval, tools, memory and token cost.
09/04/2026
→
GPT-6 Astra API Guide: Pricing, Claude/Gemini Comparison, 1.05M Context and Migration
GPT-6 Astra API pricing, 1.05M context, Responses migration and rollout, plus Claude Fable 5.1 and Gemini 3.8 Flash comparisons for coding, agents and cost.
09/04/2026
→
Gemini 3.8 Flash vs 3.7 Flash: Same Price, Different Agent Cost?
Gemini 3.8 Flash vs 3.7 Flash: compare price, 1M context, thinking levels, AI coding, AIGC, Claude/GPT selection context, agent routing, migration and token cost.
09/03/2026
→
Claude Fable 5.1 vs Mythos 5.1: Pricing, Access, Coding, and Agent Trade-offs
Claude Fable 5.1 and Mythos 5.1 share the same underlying model but differ in safeguards and access. Compare pricing, cache costs, coding benchmarks, and who should use each.
09/02/2026
→
MCP StreamableHTTPClientTransport POST Stays Pending After SSE Disconnect: Issue #2739 Reproduced
A v1 MCP StreamableHTTPClientTransport POST can remain pending until request timeout after request-scoped SSE error/EOF. XBSTACK reproduced Issue #2739 on SDK 1.29.0 and 1.30.0.
09/01/2026
→
How to Configure llms.txt v2: rel=alternate, describedby, and Markdown Discovery
llms.txt v2 adds explicit discovery for Markdown pages and llms.txt through rel=alternate and rel=describedby. This guide shows how to implement and verify both relationships.
08/31/2026
→
How to Security-Check an MCP Configuration: Secrets, Shell, Remote MCP, and Permissions
Review MCP configuration for credential exposure, shell execution, remote MCP trust, permissions, logging and retention using MCP guidance, OWASP, and Agent Security Auditor.
08/31/2026
→
How to Test an MCP Server Before Production: Read-Only Inspector Preflight
Test an MCP Server before production with read-only server/discover and list checks for protocol version, catalogs, authorization, cache hints, and modern/legacy compatibility.
08/31/2026
→
How to Add WebMCP to a Website: Chrome 149 Forms, Tool Schemas, and registerTool
Add WebMCP to a website with Chrome 149 declarative forms or document.modelContext.registerTool, then validate schemas, permissions, Lighthouse results, evals, and telemetry.
08/31/2026
→
n8n Execution Stuck Running After Success: Zombie Execution When Successful Runs Are Not Saved
n8n execution stuck running after success? On 2.36.6, a webhook can return 200 while execution_entity stays running. This A/B reproduces the cause and workaround.
08/29/2026
→
