
Claude Skills by NousResearch
github.com/NousResearchHost modded Minecraft servers (CurseForge, Modrinth).
Play Pokemon via headless emulator + RAM reads.
Manual OAuth for remote MCP servers on headless gateways.
lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.).
W&B: log ML experiments, sweeps, model registry, dashboards.
vLLM: high-throughput LLM serving, OpenAI API, quantization.
HuggingFace hf CLI: search/download/upload models, datasets.
SAM: zero-shot image segmentation via points, boxes, masks.
OBLITERATUS: abliterate LLM refusals (diff-in-means).
Pay HTTP 402 APIs via Machine Payments Protocol (MPP).
Agent payments via Stripe Link — cards, SPT, approvals.
Provision SaaS services + sync creds via Stripe Projects.
Turn an unanswerable decision into a questionnaire doc.
Shop catalog search, checkout, order tracking, returns.
Monitor blogs and RSS/Atom feeds via blogwatcher-cli tool.
Agent RAG and long-term memory with Pinecone.
Write ML papers for NeurIPS/ICML/ICLR: design→submit.
Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.
Autonomously remove your info from data-broker sites.
Control Philips Hue lights, scenes, rooms via OpenHue CLI.
AST-aware structural code search and rewrite via ast-grep.
Adversarial plan interview before implementation.
Execute plans via delegate_task subagents (2-stage review).
Deploy a Worker live, no account, via wrangler --temporary.
Record a site's XHR into a HAR, derive an HTTP client.
Versioned site deploys to GitHub/Cloudflare/Netlify Pages.
Yuanbao (元宝) groups: @mention users, query info/members.
Drive the desktop background-first; escalate on signal.
Review Kanban handoffs and route verified outcomes.
Triage an inbox: prioritize threads, draft replies safely.
Box manages cloud files, sharing, search, and metadata.
Extract cited obligations, deadlines, tasks from documents.
Create, read, edit, template, and review Word .docx files.
Turn meeting notes into cited decisions, owners, tickets.
PDF files: create, read, merge, fill, OCR, edit text.
Watch product, flight, or listing prices; alert on target.
Weekly reset: commitments, stalled work, next-week plan.
Create, read, edit Excel .xlsx workbooks and CSVs.
Watch named companies for material news; cited digests.
Ground answers and documents in cited, verifiable sources.
Inspect codebases w/ pygount: LOC, languages, ratios.
Exploratory QA of web apps: find bugs, evidence, reports.
GitHub via gh CLI: PRs, issues, reviews, repos, auth.
Read the live Hermes desktop DOM/CSS over CDP.
Parallel 4-agent cleanup of recent code changes.
Use when a fetch fails: 403/429, paywall, WAF, bot wall.