Reading
agent-trace is an observation tool designed specifically for AI agents. It can capture every tool call, prompt, and response from MCP clients like Claude Code and Cursor, helping developers debug complex workflows.
Section 01
agent-trace is an observation tool designed specifically for AI agents. It can capture every tool call, prompt, and response from MCP clients like Claude Code and Cursor, helping developers debug complex workflows.
Section 02
The design goal of agent-trace is to provide full observability of AI agent behaviors. It can:
Section 03
Modern AI agents complete complex tasks by calling external tools (such as file reading/writing, code execution, API requests). agent-trace records the parameters, execution time, and return results of each tool call, forming a complete call chain.
Section 04
It not only records the requests sent by the agent but also includes the user's original prompt and the model's complete response. This is crucial for understanding the agent's decision-making logic.
Section 05
The captured trace data can be fully replayed. Developers can reproduce the agent's entire execution process like watching a video, precisely locating the moment when a problem occurs.
Section 06
agent-trace works based on the MCP (Multi-Client Protocol) and supports:
Section 07
When an agent performs multi-step tasks (e.g., "analyze codebase, identify performance bottlenecks, generate optimization plans"), errors in any intermediate step may cause the final result to deviate from expectations. agent-trace allows developers to check each decision point step by step.
Section 08
In an enterprise environment, understanding which data the AI system accessed and which operations it performed is a basic requirement for compliance audits. agent-trace provides non-repudiable operation logs.
Keep going with more reads from the same topic.
Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.
Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.
This project explores how to use Large Language Models (LLMs) to detect logical vulnerabilities in Ethereum smart contracts, especially business logic flaws that are difficult for traditional static analysis tools to capture, such as economic manipulation and execution order errors.
This article details how to quickly deploy a virtual machine equipped with 4 NVIDIA V100 GPUs on Azure using Terraform to enable local inference of large language models with over 70B parameters. It covers a complete practical guide from infrastructure deployment, Ollama/vLLM dual-engine comparison testing, to cost optimization and actual performance data.