发现优质的 AI Agent 技能

来自 openclaw/skills 技能

**Note:** These are reference commands for manual use. This skill does not execute commands automatically — it provides examples for the user to run themselves.

Platform: PS5 Hours: ~30 Where I Left Off: Just beat Margit - Baldur's Gate 3 — need 100 hours clear

- Missing `const` — non-const widgets rebuild children even if unchanged - ListView without builder — loads all items into memory, use `.builder` - Large list itemExtent — providing fixed height enables optimizations - Image not cached — use `cached_network_image` package - setState scope — rebuilds entire widget, extract child widgets

来自 openclaw/skills 技能

Use this when the user wants viral growth, creator-style marketing, or faster message testing.

来自 openclaw/skills 技能

Download videos from almost any platform using the powerful `yt-dlp` tool.

Transcribe YouTube videos and local media files with speaker diarization via ElevenLabs.

Generate AI images and videos using RunPod public endpoints. All output is saved to `~/runpod-media/`.

A 3-agent sequential workflow that analyzes a YouTube video, writes a blog post, and generates a header image.

A 3-agent sequential workflow that researches a topic, writes a video script, and generates AI video.