Developer Tools
Turn recurring engineering problems into tools you can use directly. Lunest and Recalia now live in the dedicated Products section; this page is reserved for working utilities and developer capabilities.
⌕Search tools or use cases…AI Agent Production Check
This is not another opaque score. It sequences discovery, protocol, security, compatibility, and real-outcome verification into a practical pre-production workflow; use only the steps your system needs.
Submit only a public URL. Do not enter tokens, cookies, API keys, or login credentials.
Website / Agent Readiness
Verify whether the site can be discovered and read by agents: robots, sitemap, Markdown, llms.txt and capability-discovery signals.
MCP Endpoint Preflight
Run a read-only check of a public MCP endpoint: protocol generation, Tools, Resources, Prompts and authorization boundaries.
Agent / MCP Security
Before production, inspect secrets, shell access, remote MCP, permissions, retention and logging risks. Sensitive input stays local.
Compatibility / Known Bugs
Check observed public-server behavior plus MCP / AI Agent SDK bugs, affected versions, and fix or verification status tracked by XBSTACK.
Outcome / Side Effects
Verify real terminal state and side effects before release instead of trusting the final answer or well-formed tool calls.
Tools
AI Finance
Analyze public financial reports and company fundamentals while separating source facts, model interpretation and risk notes.
Agent Readiness Auditor
Inspect robots, sitemap, Markdown, llms.txt and optional MCP/WebMCP capabilities from an agent-readiness perspective with reviewable evidence.
MCP Inspector
Inspect public MCP Server endpoints for the 2026-07-28 stateless protocol, legacy compatibility, capabilities and tools/resources/prompts lists without invoking tools.
MCP / Agent Compatibility Tracker
Track public MCP Server protocol behavior plus MCP, Vercel AI SDK, LangGraph, and OpenAI Agents SDK issues, affected versions, and XBSTACK verification status.
Agent Security Auditor
Scan Agent/MCP configuration and code locally for shell, secret, remote MCP, permission, retention, and logging risks with evidence and remediation.
Agent Side-Effect Evaluator
Grade agents by terminal backend state and real tool effects, detecting false success, wrong state, missing/duplicate/forbidden effects, and repeated-run inconsistency.
Local AI Model Fit
Estimate whether a local model is worth testing from memory, parameter count, GGUF quantization and runtime reserve without inventing speed numbers.
Compound Calculator
Model compounding, recurring contributions, inflation and target assets with explicit assumptions.
