Design, code and verify hash-seeded long-form generative artworks before minting on platforms like Art Blocks, Highlight or 256ART.
Turns Claude into a cofounder-style brainstorming partner that writes no code, challenges your assumptions, and ends with a structured Brainstorm Brief.
An author-facing reference for how Claude Code plugins, skills, commands, agents, and hooks actually behave, with verified contracts and re-verification commands.
A four-layer project-memory playbook: search GitHub Discussions, the decision journal, and ADRs via verified GraphQL/ripgrep commands before re-investigating anything.
Turns a parsed document plus a user intent into four auditable card types — blue facts, green explanations, yellow risks, and red actions — with mandatory source citation.
A companion skill that chats with you while long background tasks run, generating scene-based "life photos" and warm messages delivered through Feishu.
A zero-trust gatekeeper skill that scans every incoming file or URL and returns a pass/reject verdict plus a compliance scan report before any document processing happens.
Writes copy-ready image prompts with the right model, quality level, and aspect ratio for your request.
An umbrella design skill with six modes: UX copy, design systems, design critique, WCAG accessibility audits, developer handoff specs, and user research synthesis.
A skill that walks Claude through configuring and launching large-scale LLM pretraining with torchtitan — 4D parallelism (FSDP2/TP/PP/CP), Float8, and distributed checkpointing from 8 to 512+ GPUs.
A hands-on guide to nanoGPT: train a tiny Shakespeare model on CPU or reproduce GPT-2 (124M) on multi-GPU, step by step.
A skill that walks you through loading, fine-tuning (LoRA/QLoRA), pretraining and deploying 20+ LLM architectures with Lightning AI's LitGPT.
Hands-on skill for installing, running and benchmarking Mamba-1/Mamba-2 linear-complexity SSMs.
Extracts journal metadata from an author-guidelines page and generates two canonical profiles — a detailed writing profile and a compact journal-matching profile.
An advisory skill that turns a medical-imaging research question into a paper-grounded shortlist of model architectures.
Reverse-engineers a researcher's publication strategy from PubMed into a CSV dataset, seven charts, and a strategy report.
Generates, audits and fixes dependency-free static HTML landing pages built for 100/100 PageSpeed with full JSON-LD markup.
Forces a probe → hypothesis → smoke → control → claim discipline when an ML training run fails, diverges, OOMs, hangs, or looks slow.
A five-step pre-flight checklist — config diff, run naming, path verification, monitoring, and ETA — plus restart/kill cleanup for long GPU training runs.
An architecture skill that audits your layers, boundaries and SOLID usage against the Dependency Rule and prescribes concrete inversions.