oneclaw

Full-stack Hermes Agent health audit: observability config, compression tuning, prompt/skill consistency, skill invocation diagnosis, system overload detection, session pattern analysis, tool-use/delegation tuning, plus a full annotated config (model/queue/IO/security), per-channel Feishu output discipline, a prompt-engineering method, and a reverse-QA regression loop. Use when skills stop triggering, model ignores loaded skill steps, context grows unexpectedly, sessions hit compression too often, long tasks get interrupted, the task queue misbehaves, or after editing any prompt/skill/config file.

Securely serve private S3 content through CloudFront using Origin Access Control (OAC). Use when you need to make S3 files publicly accessible over a CDN without making the bucket itself public — e.g. hosting static sites, media, downloads, or build artifacts. Covers the modern OAC approach (AWS-recommended), migrating legacy OAI distributions to OAC, locking down the bucket Public Access Block, enforcing HTTPS, and verifying the result. Never make an S3 bucket public directly.

用 AWS Bedrock AgentCore 云端浏览器做联网深度调研、搜索、抓网页正文。Use when 要做带引用的多源调研 / 搜资料 / 抓动态页 / 反爬站 / SPA / 需 JS 渲染的页面。本机已部署同名 MCP server，优先用它而非裸 WebFetch。NOT for 抓本机已登录态的页面(用 chrome-cdp 9222)，NOT for 纯本地文件搜索(用 grep/search_files)。

用本机已登录的 Perplexity / Gemini 做 AI 深度搜索,直接拿带引用的综合报告。当需要快速拿一个话题的多源综述、对比、"现在最好的 X 是什么"、带引用的事实核查,且想省去自己拼 web_search+抓取时使用。走 CDP 驱动本机 Chrome(9222,已登录 Pro)。NOT for 抓指定单个 URL 的正文(用 fetch_page),NOT for 需要自己控制每个来源的严谨调研(用 deep-search-protocol + agentcore-deepsearch)。

Desktop notifications for Claude Code via Notification/Stop hooks. Pops a native macOS notification showing which project, who's waiting, and what for — with click-to-focus back to your terminal. Use when setting up sound alerts, desktop notifications, or 'tell me when Claude needs me / is done'.

主动管理 context 质量与信噪比，确保决策基于可信、新鲜、相关的信息

对抗自审 — 在非平凡决策前触发魔鬼代言人式自我审查，防止幻觉和未验证断言流入输出。

本地 AI 中文双人播客生成系统（跨平台：Apple Silicon 原生 + Linux x86_64 云端）。主题一句话 → 自动调研 → 脚本 → TTS 合成 → 液体玻璃播放器 HTML。自带一男一女两位预制主持人声音，开箱即用。当用户提到播客、podcast、圆桌派、roundtable、做一期节目、生成音频、把文章变成语音、读一下这段话、有声书、做个音频节目时触发。

Abel 的浏览器自动化决策与执行指南。在 CDP MCP（精确单步）、browser-use Agent（多步探索）、WebFetch（静态读）、本地 grep（文档调研）之间做选择，并内置失败 fallback 链。

Browser automation via Chrome DevTools Protocol (CDP). Use for UI automation, form filling, data scraping, E2E testing, screenshot capture, performance auditing, and interacting with logged-in platforms. Use this instead of the browser tool when you are in Claude Code context.

Install and operate qmd (Quick Markdown) as OpenClaw's memory backend. Replaces the default SQLite memory with plain-Markdown files you can grep, diff, and version control, backed by a local better-sqlite3 vector index. Covers install, multi-agent XDG isolation, launchd auto-sync (15-min cadence), index rot detection, and native-module rebuild after Node upgrades.

Convert any OpenClaw cron job running > 25 min into the async-fork pattern. Solves the Bash-tool 1800s SIGKILL that silently kills long pipelines. Use whenever a cron job 'mysteriously disappears at 30 min', when adding a new pipeline that can exceed 25 min, or when delivery.mode: announce spams channels with started/idle stubs.

Upgrade OpenClaw to a specified version on macOS. Covers version check, proxy/network, pnpm install, config migration, launchd re-registration, and end-to-end verification. Triggered by: upgrade openclaw, update openclaw, openclaw update, openclaw upgrade.

Battle-tested OpenClaw 5.x configuration patterns — qmd memory backend, async-fork long-running cron, isolated sessions, multi-profile delivery routing, channel plugin hygiene. Use when tuning openclaw.json for production stability or when cron jobs mysteriously die at ~30 min.

Highest-accuracy web page extraction for research. Parallel three-way extraction (Trafilatura + Readability.js + VLM full-page screenshot) reconciled by an LLM. Use when you need accurate, structured content from any URL — especially login-walled, JS-heavy, or bot-blocked sites. Complements chrome-devtools skill (which is for interactive UI automation).

Use Claude Code as your autonomous coding agent. Covers task dispatch via tmux, interactive/background/loop modes, progressive task delivery, slot-machine recovery, and best practices for getting the most out of Claude Code via OpenClaw.

Configuration consistency auditing for Claude Code setups. Use when checking CLAUDE.md for contradictions, detecting stale references, finding redundant rules, or auditing configuration health after changes.

Structured research methodology for technical investigations. Use when evaluating technologies, comparing solutions, investigating incidents, or any task requiring verified, multi-source information gathering.

Guide for creating Claude Code skills from scratch. Use when building new skills, understanding skill file structure, writing effective descriptions, or improving existing skills.

Generate professional SVG architecture diagrams for GitHub README. Use when the user asks for an architecture diagram, system diagram, infra diagram, or wants to visualize project structure in a README. Also triggers on "画个架构图", "架构图", "system diagram", "infra overview", or any request to create visual documentation for a repo. Outputs a dark-theme SVG that renders natively on GitHub (no image hosting needed, works in private repos).

Chat-based AWS infrastructure assistance using AWS CLI and console context. Use for querying, auditing, and monitoring AWS resources (EC2, S3, IAM, Lambda, ECS/EKS, RDS, CloudWatch, billing, etc.), and for proposing safe changes with explicit confirmation before any write/destructive action.

Vet ClawHub skills for security and utility before installation. Use when considering installing a ClawHub skill, evaluating third-party code, or assessing whether a skill adds value over existing tools.