Machine learning

Grounds answers about GenAI/LLM/agentic AI security in the OWASP GenAI Security Project's published research -- the LLM Top 10, the Agentic Top 10, Data…

文字資料研究架構師(行銷/資管/財金的 text-as-data 前緣):把非結構文字(財報 MD&A、法說會逐字稿、消費者評論、社群輿情、客服紀錄)變成可發表的研究變數。涵蓋:語料與構念的效度論證(這些字真的量到你的構念嗎)、方法階梯選型(字典法→主題模型 LDA/STM/BERTopic→嵌入分類→LLM…

Answer a question from the personal knowledge base (LLM wiki) at ~/my-knowledge-base.

Select and run Python SDK verification with nox, Makefile targets, Ruff, mypy, pytest markers, sanity tests, type inference checks, and build checks.

Import a source folder into a knowledge-first skill by analyzing the folder with LLM assistance and generating SKILL.md plus skill.json.

PyTorch library for audio generation including text-to-music (MusicGen) and text-to-sound (AudioGen).

State-of-the-art text-to-image generation with Stable Diffusion models via HuggingFace Diffusers.

Use for "interrogate", "adversarial review", "multi-model review", "challenge this", "stress test this code", "find blind spots", or "tear this apart".

Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM.

跨版本符号迁移与二进制差分。当你有旧版本的符号/逆向结果，需要快速迁移到新版本时使用。 适用场景：内核缺 PDB 用旧版符号推导、程序更新后批量迁移函数名、应用更新后快速定位新偏移。 核心方法：用 LLM 做结构化差异比对，程序化输入输出，成本极低（200 函数 ~1 元）。…

Use for authorized security assessment of LLM applications and AI agents, including prompt injection, tool abuse, RAG exposure, memory poisoning, and model…

Compare a new local LLM against the models we run on moria and dungeon, using the repo's CRUD-app eval, and adopt it if it wins.

Create, modify, test, or package Python provider adapters under python/providers, including framework-specific dependencies, public imports, type inference…

Use when embedding meetings.

Find the right skill, agent, command, rule, checklist, or marketplace skill in this ~/.claude toolkit for the current task.

Author or review software tests and LLM/agent evals; choose test placement and mocks, diagnose flaky CI, or repair brittle suites.

This skill should be used when the user asks to review, referee, or critique an ML/AI paper for an A* venue or laboratory pre-review.

Use when auditing a codebase for semantic duplication - functions that do the same thing but have different names or implementations.

Use when working with this machine's local LLM setup (LM Studio + mlx-dspark) -- checking status, changing context window, diagnosing a reasoning hang or dead…

Knowledge base from "Solving a Million-Step LLM Task with Zero Errors" (Meyerson, Paolo, Dailey, Shahrzad, Francon, Hayes, Qiu, Hodjat, Miikkulainen —…

Use ONLY when the user's latest message explicitly contains the `/wrap-up` command.

THE single skill to publish profitable, rejection-proof skills to the Capafy marketplace end-to-end, verification-baked.

Save and retrieve memories or embeddings via the repo helpers or API.

Expert AI/ML engineer specializing in machine learning model development, deployment, and integration into production systems.