日志分析器：用 Agent Skill 做多格式日志归纳与报告
日志分析器是 SkillHub 社区 Skill（@user_8509461a/log-analyzer-new，v1.0.0，it-ops-security），无需 API Key。支持 nginx、Apache 与应用日志，自动提取错误模式、性能指标与访问趋势并生成可视化分析报告。文章说明其定位与核心能力，并给出按 skillhub.md 用 AI 安装或 skillhub install --dir 装入 Cursor、Claude Code 等 skills 目录的步骤，以及通过 Agent 发起分析的通用用法与社区来源注意事项。
Read More一个会自进化的日志分析工具：确定性规则与 AI 推理结合的 Agent Skill
介绍 SkillHub 社区 Skill @user_00c9b356/buddy-log-analyzer（v1.0.0，it-ops-security）。官方定位为日志智能分析工具，采用 mini-swe-agent 极简理念：确定性规则与 AI 推理结合。文章说明来源、requires_api_key 标签、SkillHub CLI 安装命令及各 Agent skills 目录要点，并说明公开页未附完整 SKILL.md 时如何按通用 Skill 模式在会话中使用。
Read MoreUE Log Reader: DeepSeek Harness UE Quick Log-Reading Plugin
`ue-log-reader` is a plugin attached to DeepSeek Harness, designed to simplify Unreal Engine log analysis. It parses logs in a streaming manner through model tools and provides streaming deduplication, crash extraction, statistical color-block filtering, and one-click generation of defect ticket summaries. Supporting standard UE logs and Android `adb logcat` format, it applies read-only safety checks to efficiently locate crashes and frequent issues, reducing development troubleshooting costs.
Read Moredsh-loglens: A Bounded Cursor-Aware Log Inspection Plugin for DeepSeek Harness
`dsh-loglens` is a log inspection plugin for DeepSeek Harness, designed to address the issue of agent log reads exceeding context window limits. It provides bounded windows, cursor-aware state management, and data processing via native tools, supporting file and Shell command sources. The plugin renders compact snapshots in the Web interface, including tools for listing logs, reading, raw search, and cursor reset. Installation requires configuring the `.dsh/logs.yml` file, while usage should consider data limits, security, and fingerprinting settings.
Read Moredsh-log-analyzer: A Log Analysis Tool
This plugin is designed to automate parsing of log files to extract errors, exceptions, and warnings. It supports aggregating error distributions by signature, extracting context, identifying keywords, and removing noise, without requiring a network connection. Once installed, users can quickly locate log issues using functions such as `analyze_log` and `extract_errors`. Note that excessively long logs are read from the tail by default, and users should review the source code themselves.
Read More