Jim Liu (baoyu)
- 46 skills
- 0 followers
- 23k repo stars
- 3 weeks ago last updated
- ▌ Baocut · jimliu bundleBaoCut-only operator for the installed `bcut` CLI and `.bcut` projects. Implicitly trigger only when the request explicitly names BaoCut or `bcut`, targets a `.bcut` project or BaoCut Subtitle Studio, or continues a BaoCut workflow already established in the conversation. Do not trigger solely for generic audio or video, transcription, subtitles, translation, editing, animation, review, rendering or export, FFmpeg, or another NLE. Once triggered, execute and verify the requested BaoCut transcription, subtitle or timeline, overlay, review, render, or export workflow. Being inside the BaoCut source repository is not itself a trigger; ordinary code and documentation tasks follow repository instructions unless they also operate the product or a `.bcut` project.
- ▌ Doubao Multimodal · jimliu bundleAsync audio/video understanding with Doubao-Seed (火山方舟 Ark). Handles ASR (plain / per-character timestamps / multispeaker), AST translation, speaker diarization, subtitle alignment, audio-visual captioning, video timeline JSON, technical-blog keyframe selection (with optional transcript hint), and free-form prompts over an audio or video file. Auto-uploads local files to Volcano TOS and auto-downloads remote URLs. Auto-splits videos > 20 min (≤ 50 MB/segment) and audio > 120 min (≤ 50 MB/segment) and merges results. Use when the user wants to transcribe, translate, diarize, caption, pick blog illustration frames, or ask questions about an audio/video file or URL.
- ▌ Codex App Ref Refresh · jimliu bundleRefresh the local ChatGPT app reference source from the installed macOS ChatGPT.app asar. Use when Codex needs to get the latest installed ChatGPT app code, recreate ./ref from /Applications/ChatGPT.app/Contents/Resources/app.asar, or format extracted JavaScript and CSS while excluding ref/node_modules.
- ▌ Deobfuscate Javascript · jimliu bundleReverse-engineer minified, obfuscated, packed, or bundled JavaScript into readable or typed TS/TSX, including Codex.app `./ref/webview/assets` restoration. Use for .min.js, dist bundles, Vite/Rollup/Webpack chunks, obfuscated snippets, or requests to humanify, deobfuscate, deminify, unpack, decode, make readable, rename variables, reverse engineer, 完整还原, 深度还原, 反混淆, 美化代码, 看懂这段 JS, 重命名变量, 反编译. Whole-tree restores (index.html + assets, "restore the code/whole tree") default to deep: deobfuscate if needed, smart-rename, polish, resolve npm imports, typed semantic TSX rewrite, acceptance review, import graph/ledger, and full-target quality gate draining every reachable project-local chunk. Quick/readable/快速 or lone snippets use readable naming-quality restore. Vendor data/packages become npm leaves or re-export shims, not restored package bodies.
- ▌ Diffdock · jimliu bundlePredict small-molecule binding poses with DiffDock-L (Corso et al. 2023/2024, github.com/gcorso/DiffDock) — blind diffusion docking that places a ligand into a protein pocket without a predefined search box and ranks the samples with a learned confidence model. Reach for this skill to dock a SMILES or SDF against a PDB, to generate ranked 3D poses for a small fragment library, or to get a starting pose for downstream rescoring. DiffDock predicts geometry, not affinity.
- ▌ Customize · jimliuCreate, configure, and maintain custom agent profiles and author new skills via the `repl` tool. Use when the user wants to create an agent profile, build a custom agent, modify agent capabilities, attach or detach skills/connectors on a profile, author a skill, or inspect which connectors and tools are available. Also use whenever you need the `host.agents.*` or `host.skills.*` Python SDK.
- ▌ Scvi Tools · jimliu bundleProbabilistic single-cell RNA-seq with scvi-tools — scVI for a batch-corrected latent space, scANVI for semi-supervised label transfer, and Bayesian differential expression. Reach for this skill to integrate scRNA-seq batches, embed cells for clustering, transfer annotations from a reference onto a query, or score differentially expressed genes per cluster. For spatial deconvolution / mapping use the cell2location, DestVI, or Tangram methods instead.
- ▌ PDF Explore · jimliu bundleUse this skill when the user has attached a PDF, paper, report, or other document and the answer needs content from more than one place in it: summarize the methods or any other section, compare sections, find where a topic is discussed, read a value or label off a figure or chart, or find/list/extract every instance of something across the whole document (datasets, benchmarks, citations, figures, table rows, accession numbers — including appendices). Skip it only for a single lookup of 1–4 pages quoted in your very next response — `read_file(pages=[...])` attaches pages as images that are dropped from context after one turn, so multi-section answers end up re-reading the same ranges repeatedly. Parses the PDF once in the Python kernel: `pdf_pages` (pages as persistent text), `pdf_outline` (TOC), `pdf_scan` (rank pages by relevance), `pdf_map`/`pdf_extract` (per-page summary / structured fields via parallel haiku calls). For PDF creation/manipulation, use reportlab/pypdf directly.
- ▌ Figure Style · jimliu bundlePublication-grade figure correctness and legibility rules. Load before drawing any plot and call `apply_figure_style()` — sets a role-mapped font-size ladder, outward ticks, frameless legends, and 300-dpi output. The skill is a checklist, not a house look: data fidelity (claim-titles tested against every row, excluded data never enters summaries), label economy (floor and ceiling), colour threading, chart-choice-by-data-shape, layout, and a render-then-verify QA loop (bbox collision + per-panel perceptual check). Ships helpers: focal_palette, bar_with_points, strip_with_median, end_of_line_labels, panel_letter, set_frame, panel_crops. For multi-panel figures load `figure-composer`; for whole-paper figure arc load `paper-narrative`.
- ▌ Skill Creator · jimliu bundleCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
- ▌ Self Awareness · jimliuClaude Science's own session database schema and SDK surface for introspection via host.query(). Load this when you need to query your own conversation history, token usage, cost accounting, execution log, or artifact metadata beyond what host.frames()/host.artifacts() provide — e.g. "how many tokens has this session used", "what was my last tool call", "list every file I've written", "where are messages stored", "what tables can I query", "inspect frames.context_data", or any time you're about to PRAGMA-probe the Claude Science metadata DB to discover its schema.
- ▌ Figure Composer · jimliu bundleCompose one publication-grade multi-panel figure. Entry from a one-line claim + data refs, OR from an existing figure via `derive_outline(png)`. Runs a per-figure loop: outline (12-col grid, per-panel ask + label_budget) → fan-out one sub-agent per panel (each loads `figure-style`) → tile + stamp letters → adversarial composite review with two-tier feedback (Tier-1 outline_revisions / Tier-2 per-panel violations) → regen affected panels, ≤3 rounds. Loads panel_task / compose_figure / compose_crops / composite_review_task / derive_outline into the kernel. For one standalone plot use `figure-style`; for whole-paper figure ordering use `paper-narrative`.
- ▌ Paper Narrative · jimliu bundleJudge and reshape the STORY a paper's figures tell. Input is the work itself — manuscript (or abstract) + figure deck — no hand-written brief. `derive_paper_brief(abstract, captions)` extracts pitch/vision/per-figure-claims; a handling-editor reviewer on the full deck returns hook_verdict (would Fig 1 make me send this for review?), arc (hook→mechanism→evidence→application), figure_moves (panels in the wrong figure), missing_panels (concrete analyses to RUN), kill_list, and boldest_defensible_fig1. Hands per-figure claims to `figure-composer`. Load when writing or revising a paper.
- ▌ Compute Env Setup · jimliu bundleSet up a compute environment on a remote provider so Claude Science jobs can run there. Covers direct SSH/conda hosts, Slurm clusters, container-via-bridge runners, and managed-API providers (Modal, GCP, RunPod). Use when standing up a new provider, porting an env to a different backend, adding a tool that needs its own software stack, or wiring weight caches. Triggers on "new compute provider", "set up env on", "port env to", "build GPU image", "weight cache", "compute_details", "conda env on the box", "apptainer on slurm".
- ▌ Literature Review · jimliu bundleFind, verify, and synthesize scientific literature — from "what's the seminal paper for X" through full multi-source reviews. Covers grounding claims in real retrieved sources, avoiding fabricated citations, handling retractions, and calibrating confidence to evidence strength.
- ▌ Indication Dossier · jimliu bundleGenerate a therapeutic indication dossier. Covers the patient population, epidemiology, disease biology, standard of care, regulatory precedent, and landmark clinical trials.
- ▌ Remote Compute Ssh · jimliuSubmit→wait_for_notification→harvest workflow for the user's SSH/SLURM hosts. Load once you've decided to dispatch remote.
- ▌ Interview Writing · jimliu将播客/视频访谈的分析素材转化为访谈实录整理文章。按访谈原始 顺序组织，保留问答结构的对话感，补充背景解读，让读者通过文字快速了解 访谈全貌。当用户已完成访谈内容分析（有素材包和提纲），需要撰写正式文章 时使用此 skill。
- ▌ Book Illustrator · jimliu bundle《图解 Skill —— AI 提效实战指南》专用配图技能。分析本书章节、附录或书稿内容，批量识别需要视觉辅助的位置，并行生成本书统一风格的插图提示词，批量出图，最后回写书稿。 当用户要求"为本书配图"、"给本书章节生成插图"、"补画书中插图"、"重画本书配图"、"生成章节信息图"时使用此技能。 不用于普通文章、营销图、社媒图或其他书籍的通用配图任务。
- ▌ Content Analyzer · jimliu素材深度分析技能。当用户提供文章、推文、视频文稿等内容并要求分析时使用。 适用于：(1) 理解素材核心观点和论证逻辑，(2) 批判性评估内容，(3) 为后续写作提取可用素材， (4) 分析内容的传播价值。支持长文、短内容（推文/帖子）、视频文稿等多种内容类型。
- ▌ Baoyu Weekly Report · jimliu bundle生成中文周报（Markdown 表格格式）。当用户提到"周报""写周报""本周工作""weekly report"或描述本周完成的工作内容时触发。即使用户只是随意列出这周做了什么，也应该触发这个技能来帮忙整理成周报。
- ▌ Info Digest · jimliu把任何输入的信息快速转化为简洁、易懂的中文资讯，让读者一两分钟抓住要点。 适用于新闻、报道、长文、研究报告、论文、政策文件、产品资料、官方博客、链接等各类信息。 当用户提供一段内容（尤其是英文/外文，或较长、较专业的材料），或贴出链接、标题，要求 "整理成中文资讯""写成快讯/速递/简报""用大白话讲讲""帮我把这个讲清楚""划个重点"时使用。 即使用户没明确说"写简报"，只要意图是把信息转述成给中文读者看的简短、好懂版本，也应触发。 需要时会联网检索背景、补充前因、查证事实。与深度分析（content-analyzer）不同， 本技能产出的是简短、可一两分钟读完的纯文本资讯，不是长篇结构化报告。
- ▌ Outliner · jimliu科技专栏提纲生成技能。当用户提供素材（文章、推文、视频文稿等） 要求生成写作提纲时使用。生成 3-5 个差异化大纲方案（优先包含故事驱动型）， 每个方案保存独立文件供用户选择。
- ▌ Baoyu Design · jimliu bundleCreate polished design artifacts as self-contained HTML: UI mockups, interactive prototypes, wireframes, landing pages, dashboards, app screens, mobile apps, slide decks (a.k.a. PPT / PowerPoint presentations), documents and résumés, animations, 3D objects, research reports, HTML email, diagrams, fliers, and visual explorations. Use whenever the user asks to design, mock up, prototype, wireframe, visualize, explore, or make a PPT/deck for an interface, product screen, user flow, content layout, visual artifact, or pitch/deck concept, even if they do not say "design". Also use to export a deck built with this skill to PowerPoint (PPT/PPTX) — but only decks authored here (deck-stage / this skill's slide-structured HTML), NOT arbitrary HTML, so confirm the target is such a deck first. Also use for setting up, importing, or authoring reusable design systems, UI kits, brand tokens, or component libraries. Harness-agnostic for Claude Code, Cursor, Codex Agent, and similar file-capable agents.
- ▌ Baoyu Image Gen · jimliu bundleGenerates images using multiple AI providers including OpenAI, Google, Azure, and others. Supports text-to-image, reference images, aspect ratios, and batch generation from prompt files.
- ▌ Baoyu Post To X · jimliu bundlePosts text, images, videos, and long-form Markdown articles to X (Twitter) using a real Chrome browser with multiple automation modes.
- ▌ Baoyu Translate · jimliu bundleTranslates text between languages with three modes: quick, normal, and refined. Supports custom glossaries, audience targeting, and style presets.
- ▌ Baoyu Slide Deck · jimliu bundleGenerates professional slide deck images from content, creating outlines with style instructions and individual slide images.
- ▌ Baoyu Xhs Images · jimliu bundleBreaks down complex content into eye-catching image card series with 12 visual styles, 8 layouts, and 3 color palettes, optimized for social media engagement.
- ▌ Baoyu Cover Image · jimliu bundleGenerates article cover images with customizable type, palette, rendering, text, and mood dimensions, supporting multiple aspect ratios and backends.
- ▌ Baoyu Infographic · jimliu bundleGenerates professional infographics by combining 21 layout types with 22 visual styles, analyzing content to recommend optimal combinations, and producing publication-ready raster images.
- ▌ Baoyu Post To Weibo · jimliu bundlePosts text, images, videos, and long-form Markdown articles to Weibo via a real Chrome browser, bypassing anti-bot detection.
- ▌ Baoyu Compress Image · jimliu bundleCompresses images to WebP (default) or PNG with automatic tool selection, reducing file size while preserving quality.
- ▌ Baoyu Post To Wechat · jimliu bundlePosts content to WeChat Official Account via API or Chrome CDP, supporting articles with HTML/markdown/plain text and image-text posts with multiple images.
- ▌ Baoyu Wechat Summary · jimliu bundleSummarizes WeChat group chat highlights into a structured digest using the local wx-cli binary, with per-group history, user profiles, and privacy guardrails.
- ▌ Baoyu Format Markdown · jimliu bundleTransforms plain text or markdown into well-structured, reader-friendly markdown with frontmatter, titles, summaries, headings, bold, lists, and code blocks.
- ▌ Baoyu Url To Markdown · jimliu bundleFetch any URL and convert it to clean markdown using a Chrome CDP-based CLI with site-specific adapters for X/Twitter, YouTube, Hacker News, and generic pages.
- ▌ Release Skills · jimliuAutomates project releases with version bumping, multi-language changelog generation, and GitHub Release creation. Supports Node.js, Python, Rust, and generic projects.
- ▌ Baoyu Electron Extract · jimliu bundleExtracts resources and JavaScript from installed Electron apps by unpacking `.asar` bundles, restoring original sources from source maps when available, and formatting minified code with Prettier.
- ▌ Baoyu Markdown To HTML · jimliu bundleConverts Markdown to styled HTML with WeChat-compatible themes, supporting code highlighting, math, Mermaid diagrams, PlantUML, footnotes, alerts, and optional bottom citations for external links.
- ▌ Baoyu Danger Gemini Web · jimliu bundleGenerates images and text via reverse-engineered Gemini Web API, supporting reference images and multi-turn conversations.
- ▌ Baoyu Youtube Transcript · jimliu bundleDownloads YouTube video transcripts/subtitles and cover images by URL or video ID, with support for multiple languages, translation, chapters, and speaker identification.
- ▌ Baoyu Article Illustrator · jimliu bundleAnalyzes article structure, identifies positions requiring visual aids, and generates illustrations with consistent type, style, and palette.
- ▌ Baoyu Danger X To Markdown · jimliu bundleConverts X (Twitter) tweets, threads, and articles to markdown with YAML front matter using a reverse-engineered API.
- ▌ Baoyu Comic · jimliu bundleCreates original educational comics with flexible art styles and tones, supporting detailed panel layouts and batch image generation.
- ▌ Baoyu Diagram · jimliu bundleCreate professional, dark-themed SVG diagrams of any type — architecture, flowcharts, sequence, structural, mind maps, timelines, and more.