Avatar 🍥

KnightLi Blog

记录并分享日常

  1. Home
  2. About
  3. Archives
  4. Search
  5. Links
    1. Dark Mode

Archives

2026 58
2025 23
2024 5
2023 9
2022 30
2021 5
2020 8

Categories

Technical Docs AI Tools Hardware Operations Development Tools AI Industry AI工具 Blockchain

Tags

Ollama Ubuntu Local LLM Gemma 4 AI Agent Python Agent-Skills Hugo Local LLMs Nginx OpenClaw API GGUF Linux Windows Certbot Codex Hugging Face Llama.cpp Openwrt Pinout Quantization VS Code WordPress Agent Harness Ai-Practice Anthropic Atx-Power-Supply Backblaze-Drive-Stats Chip-Overview
AI Tools

What Is OpenHarness: What This Open Source Agent Harness Can Do

Based on the official HKUDS/OpenHarness repository and README, this article summarizes OpenHarness's positioning, core capabilities, ohmo's personal-assistant features, and the scenarios it fits best.

2026-04-12
6 minute read
中文简体 中文繁體
AI Tools

Getting Started with Playwright CLI: Installation, Skills, Sessions, and Essential Commands

Based on the latest microsoft/playwright-cli README, this guide walks through Playwright CLI's positioning, installation, skills workflow, session management, monitoring dashboard, and essential commands.

2026-04-12
7 minute read
中文简体 中文繁體
AI Tools

What Is Hermes Agent: Overview, Strengths, Getting Started, and How It Compares to OpenClaw

A practical introduction to Nous Research's Hermes Agent: what it is, where it stands out, how to get started, and how it differs from OpenClaw in positioning and user experience.

2026-04-12
7 minute read
中文简体 中文繁體
AI工具

OpenClaw Dreaming: Machines Start Dreaming While Humans Lose Sleep

OpenClaw introduced Dreaming, a memory consolidation system modeled on light sleep, deep sleep, and REM to help agents retain signal and discard noise.

2026-04-12
4 minute read
中文简体 中文繁體
AI Tools

How to Use llama-quantize for GGUF Models

A short introduction to what llama-quantize does, its basic commands, common options, and the tradeoffs between model size, speed, and quality.

2026-04-12
2 minute read
中文简体 中文繁體
AI Tools

How to Get GGUF Models from Hugging Face with llama.cpp

A short guide to downloading GGUF models with llama.cpp from Hugging Face, switching compatible endpoints, and converting non-GGUF formats.

2026-04-12
1 minute read
中文简体 中文繁體
AI Tools

Codex Usage and Quota Check

Use a small Python script to read credentials from `auth.json`, call ChatGPT's `/backend-api/wham/usage` endpoint, and inspect remaining Codex quota plus reset times.

2026-04-12
10 minute read
中文简体 中文繁體
AI Tools

What Does `it` Mean in Gemma-4-31B-it

A brief explanation of what `it` and `31B` mean in Gemma-4-31B-it, and why `it` is usually the right choice for chat use.

2026-04-11
1 minute read
中文简体 中文繁體
AI Tools

Choosing Llama GGUF Quantization on Hugging Face: Practical Advice from Q8 to Q2

A practical way to understand GGUF quantization levels and choose between Q8, Q6, Q5, Q4, Q3, and Q2 based on hardware limits.

2026-04-11
2 minute read
中文简体 中文繁體
AI Tools

How to Access a Local Ollama API Over LAN on Windows

Expose Ollama API to your local network on Windows by setting the host, allowing firewall ports, and verifying with curl.

2026-04-11
1 minute read
中文简体 中文繁體
Hardware

Common USB PD Decoy Chips: CH224K vs HUSB238 vs HUSB237 vs IP2721 vs XSP

A quick comparison of CH224K, HUSB238, HUSB237, IP2721, and XSP series decoy chips for USB PD power design.

2026-04-11
1 minute read
中文简体 中文繁體
AI Tools

What Models Power fnOS AI Photos: Face, Object, and Semantic Search Stack

A practical breakdown of the fnOS AI photo stack, covering face recognition, object detection, semantic search, and hardware acceleration.

2026-04-11
1 minute read
中文简体 中文繁體
Operations

go2rtc with Xiaomi Camera RTSP: Feed NVR, HomeKit, and Frigate

A practical setup note for pulling Xiaomi camera RTSP via go2rtc and using it across NVR, HomeKit, and Frigate.

2026-04-11
1 minute read
中文简体 中文繁體
AI Tools

Gemma 4 Local Runtime Guide: From One-Command Start to Dev Integration

A concise guide to main local runtime paths for Gemma 4, including Ollama, LM Studio, llama.cpp, and developer-oriented integration.

2026-04-10
2 minute read
中文简体 中文繁體
AI Tools

Drop MCP? Why CLI Is Becoming the Default Tool Layer for Agents

Across cost, reliability, training distribution, and security model, here is why more agent workflows are returning to CLI-first.

2026-04-10
3 minute read
中文简体 中文繁體
AI Tools

PersonaPlex Quick Guide: Full-Duplex Conversational Speech with Persona and Voice Control

A concise guide to PersonaPlex capabilities, setup, and prompting, including server launch, offline evaluation, and role/voice control.

2026-04-10
2 minute read
中文简体 中文繁體
AI Tools

Anthropic's Harness Direction: Agent Infrastructure Is Becoming an Agent OS

A concise breakdown of Anthropic's latest practice across session, harness, and sandbox, and why agent architecture is moving toward stable abstractions with recoverable execution.

2026-04-10
2 minute read
中文简体 中文繁體
AI Tools

OpenClaw and Agent Harness: Why It Looks Like AGI

A harness-based view of OpenClaw: the model remains the core, while autonomy comes from the engineering combination of memory, tools, triggers, and execution loops.

2026-04-10
2 minute read
中文简体 中文繁體
AI Tools

Sharing an Agent Skill for E-commerce Product Image Cutout and Standardization

An overview of the product-cutout-normalize Agent Skill, including its purpose, usage, parameters, and the full source code for SKILL.md and scripts/run_pipeline.py.

2026-04-09
10 minute read
中文简体 中文繁體
AI Tools

How to Use Google Nano Banana for Image Cutouts

Based on a practical Python example, this article explains how to use Google Nano Banana for product-image background removal while preserving the full source code.

2026-04-09
8 minute read
中文简体 中文繁體
AI Tools

What are Ollama cloud models and how do you use them

A brief explanation of what Ollama cloud models are, how they differ from local models, and how to use them from the command line or via API.

2026-04-09
2 minute read
中文简体 中文繁體
Operations

Windows Task Manager data stops updating: the refresh speed is usually set to Paused

If CPU, memory, or network data in Windows Task Manager stops updating for a long time, a common cause is that the refresh speed was accidentally set to Paused.

2026-04-09
2 minute read
中文简体 中文繁體
AI Tools

How to Download a GGUF Model from Hugging Face and Import It into Ollama

If a model does not have a ready-made Ollama version, you can download the GGUF file from Hugging Face and import it into Ollama with a Modelfile.

2026-04-09
3 minute read
中文简体 中文繁體
AI Tools

How to Troubleshoot Slow `ollama pull` Model Downloads

When `ollama pull` is slow, times out frequently, or gets interrupted, it helps to identify the real download path first and then troubleshoot the redirected object storage domain.

2026-04-09
4 minute read
中文简体 中文繁體
AI Industry

Anthropic and OpenClaw Timeline: The Full Sequence of Events

Based on public information from January to April 2026, this article outlines the process, key milestones, and current status of the Anthropic-OpenClaw developments.

2026-04-08
3 minute read
中文简体 中文繁體
AI Tools

Gemma 4 on Raspberry Pi 5: It Works, But Responses Are Slow

From LM Studio CLI deployment to LAN access and editor integration, this article benchmarks how usable Gemma 4 is on Raspberry Pi 5.

2026-04-08
3 minute read
中文简体 中文繁體
AI Tools

Connect OpenClaw to Local Gemma 4: Complete Setup Guide

From starting the Ollama API to configuring OpenClaw, this guide walks you through connecting local Gemma 4 end to end.

2026-04-08
2 minute read
中文简体 中文繁體
AI Tools

How to Run Gemma 4 on a Laptop: 5-Minute Local Setup Guide

Run Gemma 4 quickly on Mac, Windows, and Linux laptops with Ollama, plus model selection and performance tips.

2026-04-08
2 minute read
中文简体 中文繁體
AI Tools

How to Install and Run Gemma 4 on Android: Complete Getting-Started Guide

From APK installation and model download to multimodal usage, this guide walks you through running Gemma 4 on Android.

2026-04-08
3 minute read
中文简体 中文繁體
Hardware

Memory Die Identification Guide: How to Read Samsung, Micron, and SK hynix Codes

A beginner's guide to memory die identification using Samsung, Micron, and SK hynix examples, including common markings, code fields, and practical boundaries.

2026-04-06
3 minute read
中文简体 中文繁體
AI Tools

Using Agent Skills to Replace Copilot's 'Generate Commit Message' in VS Code

Use the local `git-commit-push-zh` skill to replace VS Code's commit message generation and keep a standardized commit/push workflow.

2026-04-06
2 minute read
中文简体 中文繁體
AI Tools

How to Check Whether an Ollama Model Is Loaded on GPU

Use `ollama ps` to quickly verify whether a model is running on GPU, CPU, or a CPU/GPU mixed memory setup, and learn how to read the `PROCESSOR` column.

2026-04-06
1 minute read
中文简体 中文繁體
AI Tools

Using AI Agent Skills to Maintain a Hugo Multilingual Blog

A reusable deep dive into the sync-post-translations skill: goals, rules, configuration code, usage examples, and quality checklist.

2026-04-06
3 minute read
中文简体 中文繁體
AI Tools

Ollama Default Model Storage Path and Migration Guide (Avoid Filling Up C Drive)

A practical guide to Ollama's default model directories on Windows, macOS, and Linux, plus executable Windows steps to move models off the system drive.

2026-04-06
2 minute read
中文简体 中文繁體
AI Tools

Completely Uninstall Ollama on Linux (Including Leftover Cleanup)

A practical, executable guide to fully remove Ollama on Linux by stopping and removing its service, binary, model files, and system user/group.

2026-04-06
2 minute read
中文简体 中文繁體
AI Tools

LLM Quantization Explained: How to Choose FP16, Q8, Q5, Q4, or Q2

A practical guide to LLM quantization, common format differences, and VRAM-based model selection to balance quality, speed, and resource cost.

2026-04-05
2 minute read
中文简体 中文繁體
AI Tools

Google Gemma 4 Model Comparison: How to Choose Between 2B/4B/26B/31B

A structured comparison of Gemma 4's 2B, 4B, 26B, and 31B variants, including performance positioning, VRAM requirements, real-world scenarios, and model selection guidance.

2026-04-05
3 minute read
中文简体 中文繁體
AI Tools

Analyzing Anthropic's docx Agent Skill: Features, Code Structure, Usage, and Caveats

Based on SKILL.md and the supporting scripts under Anthropic's skills/docx, this post breaks down the docx skill's capability boundaries, code structure, practical workflow, and common pitfalls.

2026-04-04
6 minute read
中文简体 中文繁體
Operations

Two Ways to Remotely Access Feiniu NAS and Their Comparison

A practical comparison of two remote-access options for Feiniu NAS: direct public IP access and FN Connect, including setup points, pros/cons, and usage recommendations.

2026-04-04
3 minute read
中文简体 中文繁體
Hardware

JMicron Chip Top Mark and Part Number Format

A practical guide to JMicron chip top-mark and part-number coding for identifying vendor, package, temperature grade, mask version, and IC revision.

2026-04-04
2 minute read
中文简体 中文繁體
Development Tools

CH347 Resource Guide: Drivers, Tools, and SPI Flash Programming

A beginner-friendly CH347 resource roundup: official page, Windows drivers, SPI Flash programming tools, and troubleshooting tips.

2026-04-03
2 minute read
中文简体 中文繁體
Operations

Automatically Renew Let's Encrypt Certificates on Ubuntu (Certbot + Nginx)

A practical Certbot auto-renew guide with a cron example, Nginx pre/post hooks, renewal verification, and troubleshooting tips.

2026-04-03
2 minute read
中文简体 中文繁體
Development Tools

FFmpeg `-map` Explained: Precisely Select Video, Audio, and Subtitle Streams

A practical guide to FFmpeg `-map`: stream naming, multi-input composition, negative mapping, and optional mapping.

2026-04-02
3 minute read
中文简体 中文繁體
Hardware

Home Printer Buying Guide

A practical guide to choosing a home printer based on real usage: laser vs. inkjet, connectivity options, duplex printing, and paper tray setup.

2026-04-01
3 minute read
中文简体 中文繁體
Development Tools

How to Troubleshoot High VS Code CPU Usage Caused by Extensions

A practical VS Code high-CPU troubleshooting guide: use Start Extension Bisect to quickly isolate problematic extensions, then optimize search and file watching settings.

2026-04-01
2 minute read
中文简体 中文繁體
Technical Docs

rsync --delete Explained and Practical Directory Cleanup

Learn what rsync --delete does, common usage patterns, and key precautions, including a practical way to quickly clean a target directory using an empty source directory.

2026-03-29
2 minute read
中文简体 中文繁體
Technical Docs

How Git Tracks File Executable Permission (+x)

Learn how Git tracks executable permissions on Linux, with practical commands to add, remove, and verify the executable bit.

2026-03-29
1 minute read
中文简体 中文繁體
Technical Docs

Hugo Shortcode Practice: bundle-file (Keep Multilingual Files and Attachments in One Page Bundle)

Use a custom bundle-file shortcode to automatically fall back to other translations when the current language resource is missing, and output a unified attachment link.

2026-03-29
2 minute read
中文简体 中文繁體
Technical Docs

How to Create and Use Skills: SKILL.md Specification and Practical Principles

A structured guide to SKILL.md standard structure and four core principles for writing high-quality Skills.

2026-03-28
3 minute read
中文简体 中文繁體
Technical Docs

IEEE 802.3af/at/bt (PoE) Differences Explained

A clear guide to IEEE 802.3af, 802.3at, and 802.3bt across power budget, wire pairs, and typical use cases.

2026-03-28
2 minute read
中文简体 中文繁體
1 2 3
© 2022 - 2026 KnightLi Blog
记录并分享
Built with Hugo
Theme Stack designed by Jimmy