Latest Agent Skills

Analyzing session replays, extracting persona-based behavioral patterns, and storytelling UX issues. Reads the 'why' from real user operation logs. Works with Field/Echo for persona validation.

Generating slides via Marp, reveal.js, or Slidev, designing narrative arcs, and optimizing conference talks with WPM-calibrated timing. Use when creating or pacing presentations.

Generating, updating, auditing, and sync-repairing project-specific Claude Code skills from the repo stack and conventions. Use when authoring project-local skills (Micro or Full).

Designing multi-tenant architectures with tenant isolation strategies, RLS, routing, and scale design for SaaS. Use when designing multi-tenant SaaS systems or tenant isolation.

Investigating bugs via root cause analysis, reproduction steps, and impact assessment. Investigation-only — finds why bugs occur and where to fix them, no code. Use when a bug needs RCA before a fix.

Orchestrating multi-session parallel execution via Claude Code Agent Teams API and Codex CLI Subagents — launch, manage, coordinate concurrent tasks. Use when parallel work is needed.

Defining KPIs, tracking events, and dashboards: North Star Metric, funnel and cohort analysis, test-intelligence views. GA4/Amplitude/Mixpanel/PostHog. Use when metrics design is needed.

Consultant for NotebookLM steering prompt design. Optimizes Audio/Video/Slide/Infographic output quality through source preparation, prompt engineering, and Custom Goals persona design.

Orchestrating specialist AI agent teams as a meta-coordinator: decomposes requests into minimum viable chains, spawns each as an independent session, drives to final output. For multi-domain tasks.

Converting document formats (Markdown/Word/Excel/PDF/HTML). Converts specs from Scribe and reports from Harvest into distributable formats; generates reusable conversion scripts. Use when converting documents, building accessibility-compliant PDFs, or creating Pandoc/LibreOffice pipelines.

Proposing, configuring, debugging, and maintaining Claude Code hooks (PreToolUse/PostToolUse/Stop and other lifecycle events). Use for workflow automation or quality gates via hooks.

Designing, optimizing, and auditing repository structure: directory design, docs/ layout (PRD, specs, ADR), test/script organization, anti-pattern detection. Use for repo layout or migration planning.

Use the Figma MCP server to fetch design context, screenshots, variables, and assets from Figma, and to translate Figma nodes into production code. Trigger when a task involves Figma URLs, node IDs, design-to-code implementation, or Figma MCP setup and troubleshooting.

Building engineer self-branding by turning technical contributions into a professional brand. Use for GitHub/LinkedIn/blog/conference positioning or content strategy.

Engineering privacy and data governance: PII detection, data flow mapping, consent patterns, GDPR/CCPA-compliant implementation, DPIA. Use when privacy-by-design is needed.

Search arXiv papers by keyword, author, category, or ID.

Specialist orchestrating codex exec / Antigravity CLI through dual paradigms — COMPETE (multi-variant comparison, select best) and COLLABORATE (decompose tasks across engines, integrate). Supports Solo/Team/Quick execution modes.

Use only when the user explicitly asks to stage, commit, push, and open a GitHub pull request in one flow using the GitHub CLI (`gh`).

Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like \"the xlsx in my downloads\") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.

Game audio generation agent. Produces code (Python/JS/TS/Shell) for SFX, BGM, Voice, Ambient, and UI sounds using ElevenLabs/Stable Audio/MusicGen/Suno/OpenAI TTS/JSFXR. Handles LUFS normalization and middleware integration.

Converting repository changes into detailed learning documents. Use when turning diffs into teaching materials, recording design decisions, or creating onboarding materials for new members.

Use when the user asks to generate, remix, poll, list, download, or delete Sora videos via OpenAI\u2019s video API using the bundled CLI (`scripts/sora.py`), including requests like \u201cgenerate AI video,\u201d \u201cSora,\u201d \u201cvideo remix,\u201d \u201cdownload video/thumbnail/spritesheet,\u201d and batch video generation; requires `OPENAI_API_KEY` and Sora API access.

Use this skill any time a .pptx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates, layouts, speaker notes, or comments. Trigger whenever the user mentions \"deck,\" \"slides,\" \"presentation,\" or references a .pptx filename, regardless of what they plan to do with the content afterward. If a .pptx file needs to be opened, created, or touched, use this skill.

Designing GitHub Actions workflows in depth: trigger strategy, security hardening, performance optimization, PR automation, and Reusable Workflow design.

Enumerating failure modes via pre-mortem analysis. Systematically identifies failure scenarios for plans, designs, and features, scoring them with RPN/AP. Does not write code.

Curating cross-agent knowledge and institutional memory: extracts patterns from agent journals into METAPATTERNS.md, detects knowledge decay, propagates best practices. Use for memory curation.

Domain knowledge agent for Japanese income tax filing (kakutei shinkoku). Guides income classification, deduction optimization, tax calculation, and filing for freelancers, sole proprietors, and side-business earners. Does not write code.

Comprehending and investigating codebases: structure mapping, feature discovery, data flow tracing for 'does X exist?' or 'how does Y work?'. Includes a conversational ask mode. Does not write code.

Simulating business strategy via short/mid/long-term scenario planning from financial, market, and competitive data. Applies SWOT/PESTLE/Porter, KPI forecasting, roadmaps. Does not write code.

Managing dependencies, CI/CD optimization, Docker configuration, and operational observability (logging/alerting/health checks). Use for build errors or dev environment issues.

Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation.

Proposes exactly one personal side-project idea per invocation, sized to a 1-3 day MVP. Targets CLI, automation, LLM, DX, productivity, and data-viz angles; avoids clichés like TODO apps, weather apps, and pomodoro timers. Output is an 8-section brief including a ready-to-paste coding-agent prompt. Use for morning/daily idea rituals and weekend-hack ideation. Don't use for existing-product feature proposals (Spark), dialogue brainstorming (Riff), or prototype implementation (Forge).

AI 3D model generation agent. Generates text-to-3D and image-to-3D code (Python/JS/OpenSCAD) using Meshy, Tripo, Hunyuan3D, Rodin, Sloyd, and Stability APIs. Handles game pipeline integration, LOD, retopology, UV, and QC validation.

Casting personas: rapid generation from diverse inputs, registry-based persistence and lifecycle, data-driven evolution, inter-agent sync. Not for UI walkthroughs (Echo) or user research (Field).

Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.

将一篇经管实证论文摄取为实证研究 wiki：论文卡片 + 变量 + 数据 + 模型 + 机制 + 识别 + 稳健性 + 异质性 + 表格线索

基于已摄取文献和本地数据，生成经管实证研究设计：问题、机制、变量、模型、识别、机制、异质性、稳健性

将一篇理论建模论文摄取为理论研究 wiki：论文卡片 + 假设/原语 + 命题/定理 + 解概念 + 可检验推论，并与实证层在同一张 graph 上接桥

LaTeX 编译 → PDF：latexmk 编译 + 自动修复 + 页数/匿名/字体/[UNCONFIRMED] 检查 + 提交清单

汇总某个经管实证变量在文献中的测算口径、数据来源、模型角色和项目可用性

从 PAPER_PLAN 起草 LaTeX 论文：逐 section 从 wiki 取材撰写 + 生成 figures/tables + BibTeX 验证 + de-AI polish

每日从 arXiv 拉取新论文，过滤相关性，auto-ingest 高优先级论文，检测 SOTA 更新

将实证研究设计转成 Stata 执行计划或 do 文件骨架，包含数据合并、变量构造、描述统计、回归和稳健性

从 claim graph 编译论文大纲：编译 evidence map → 叙事结构 → 章节计划 + figure plan + citation plan，Review LLM review 必选

查看所有运行中实验的状态，可选自动收集已完成实验并推进流水线

Claim-driven 实验设计：界定目标 claims → 设计实验块（baseline/validation/ablation/robustness）→ 构建执行顺序 → 可选 Review LLM review → 写入 wiki

端到端研究编排器：idea 发现 → 实验设计 → 执行 → 判决 → 论文撰写，带人工门控和状态恢复

解析审稿意见 → 原子化 concerns (Rvx-Cy) → 映射到 wiki claims → 检查 evidence → Review LLM stress-test → 生成 rebuttal

实验判决门：Review LLM 独立评判实验结果 → 4 种判决路径 → 自动更新 claims confidence、ideas status、graph edges

基于 anchor 论文、topic 关键词或当前 wiki 状态，产出一份排好序的候选论文 shortlist，供用户或上游 skill 决定是否进一步 `/ingest`。当用户问 "接下来该读什么"、"找和这篇相似的论文"、"推荐相关工作"、"这个方向周围有什么" 时触发；`/ingest --discover` 也会内部调用本 skill。本身不 ingest，只提出候选。

用领域基础知识填充 wiki/foundations/，避免后续 /ingest 为教科书材料创建重复的 concept 页面

多源 novelty 验证：WebSearch + Semantic Scholar + wiki + Review LLM cross-verify，输出 novelty 评分与建议

实验执行全流程：准备代码 → 部署运行 → 监控状态 → 收集结果，支持三种运行模式

从 wiki 知识生成论文 Related Work 章节：主题分组 → 叙事结构 → LaTeX 输出，遵循 citation-verification 和 academic-writing

通用跨模型审查：Review LLM 对任意研究制品进行独立评审，输出结构化评分、wiki 实体映射与改进建议

通用多轮迭代改进：对任意研究制品反复调用 /review → 解析反馈 → 修复 → 更新 wiki，直到达标

把一篇论文 ingest 进 wiki —— 建立 papers + concepts + people + claims 页面，并完成所有双向交叉引用与 graph edge。当用户说 "ingest"、"加入这篇论文"、丢 `.pdf` / `.tex` / arXiv URL 或要求把论文折叠进知识库时触发。

多阶段研究 idea 生成管道：景观扫描 → 双模型脑暴 → 初筛 → 深度验证 → 写入 wiki

交互式 API key 配置引导 — 检测当前 .env 状态，逐步引导配置 Semantic Scholar、DeepXiv 和 Review LLM

按 scope 重置 wiki 到干净 scaffold（wiki / raw / log / checkpoints / all）。适用于开发迭代或 setup 失败后的无痛重启。