Comentaris

hermes-agent-operator

hermeum

Scaffold a complete, commented HermesAgent custom resource YAML manifest. Reads the HermesAgent CRD schema from the live cluster (or falls back to the repo CRD file) so the output always matches the deployed spec version.

evaluate-cluster

stas00

Evaluates an ML GPU cluster for a cloud trial or acceptance test: environment dump, isolated newest PyTorch, matmul FLOPS (MAMF/MSMF) on every GPU while the others compute, intra-node all-reduce bandwidth and per-call latency of every collective, the same inter-node on every node you were given (omit those sections if there is only one node), fio on local disk and shared FS, dated markdown report. Use when the user asks to evaluate a cluster, kick the tires on trial nodes, run cluster acceptance, or measure GPU/network/storage. Canonical copy: https://github.com/stas00/ml-engineering/blob/master/skills/evaluate-cluster/SKILL.md

app-coding-standards

Expensify

Provides coding standards for the Expensify App's React Native code, including performance patterns, consistency rules, and clean React architecture. Use when writing, modifying, or reviewing App code.

aggg2-persona

youlianvr

Use for ANY communication with the user — how to speak and respond: when to explain simply and step by step, when briefly, how to react to a difficult situation, how to verify your own answers, how to reduce the interlocutor's load, how not to pass guesses off as facts. Covers: tone and style (calm, human, no sugarcoating), adaptation to the interlocutor (pace, level, mood), "explain step by step" mode, how to explain complex things (audience → layers → analogy → comprehension check), visual presentation (tables, hierarchy, highlighting, diagrams), importance of no hallucinations (fact checking, "verify" instead of guessing), testing answers through a beginner's eyes, answer structure (result first, next step last). Not for techniques (web-research-camoufox, db-first-search) — this skill is about HOW to speak.

coding-style

youlianvr

"Core principles for clean, maintainable, minimal code generation. Merged 2026-08-09 from coding-style.md + andrej-karpathy-skills.md + karpathy-coder + minimalist; extended 2026-08-15 with reviewed ponytail mechanics (DietrichGebert/ponytail, MIT) and the code-simplifier over-simplification guard (Anthropic claude-plugins-official). Enforces Karpathy's 4 principles (think before coding, simplicity first, surgical changes, goal-driven execution), the Minimalist Efficiency Ladder (YAGNI → reuse → stdlib → native → existing deps → one-liner → minimum code), clean-deletion rules, lazy-senior intensity levels (lite/full/ultra), trade-off markers, compressed output format, and a clarity-over-brevity guard against over-simplification. Use when writing, reviewing, or committing code; when the user asks for efficient code, fewer dependencies, YAGNI, 'don't overcomplicate', a 'karpathy check', or before claiming a diff is good."

reviewing-readability

streamlit

Evaluates comments, docstrings, and naming in code for readability by a developer new to the codebase — is the documentation clear and concise, and is non-obvious logic documented? Produces findings with concrete proposed rewrites (or additions where documentation is missing); the caller decides whether to apply them or present them as feedback. Use when reviewing a PR, branch, or set of changes for comment quality, naming clarity, or documentation readability. For a PR's title and description, use reviewing-pr-description instead.

sharing-pr-agent-artifacts

streamlit

Uploads agent-generated artifacts (specs, plans, learnings) to the streamlit.wiki for sharing via PR comments. Use when you have agent artifacts to share with reviewers.

ln-514-regression-checker

levnikolaevich

Worker that runs existing tests to catch regressions. Auto-detects framework, reports pass/fail. No status changes or task creation.

cm

Dicklesworthstone

"CASS Memory System - procedural memory for AI coding agents. Three-layer cognitive architecture with confidence decay, anti-pattern learning, cross-agent knowledge transfer, trauma guard safety system. Bun/TypeScript CLI."

scgpt

HughYau

Embed and annotate single-cell expression data with scGPT, a foundation model for single-cell biology. Use this skill when: (1) Producing cell embeddings from an AnnData for clustering/integration, (2) Zero-shot or fine-tuned cell-type annotation, (3) Gene-level representation for perturbation/GRN tasks. For probabilistic single-cell models (scVI etc.), use the scvi-tools library.

Ai Code Generation

omer-metin

ln-610-code-comments-auditor

levnikolaevich

Audit code comments and docstrings quality across 6 categories (WHY-not-WHAT, Density, Forbidden Content, Docstrings, Actuality, Legacy). Use when code needs comment review, after major refactoring, or as part of ln-100-documents-pipeline. Outputs Compliance Score X/10 per category + Findings + Recommended Actions.

gene-database

jaechang-hits

"Query NCBI Gene via E-utilities for curated gene records across 1M+ taxa. Retrieve official gene symbols, aliases, RefSeq accessions, summary descriptions, genomic coordinates, GO annotations, and interaction data. Use for gene ID resolution, cross-species queries, and gene function summaries. For sequence retrieval use Ensembl; for expression data use geo-database."

encode-database

jaechang-hits

"Query the ENCODE Portal REST API for regulatory genomics data: TF ChIP-seq experiments, ATAC-seq/DNase-seq accessibility peaks, histone mark tracks, and RNA-seq datasets across 1000+ cell types and tissues. Search experiments by assay, biosample, or target protein; download BED/bigWig files; retrieve candidate cis-regulatory elements (cCREs) from ENCODE SCREEN by genomic region or gene. Use for finding regulatory tracks to annotate variants, identifying open chromatin in a cell type of interest, and downloading peak files for ChIP-seq or ATAC-seq analysis. For regulatory variant scoring use regulomedb-database; for GWAS associations use gwas-database."

bcftools-variant-manipulation

jaechang-hits

"Command-line toolkit for VCF/BCF variant file manipulation. Filter, merge, annotate, query, normalize, and compute statistics on variant call files. Essential for post-variant-calling pipelines: quality filtering, multi-sample merging, rsID annotation, and genotype extraction. Companion to samtools in the HTSlib ecosystem. Use GATK instead for complex indel realignment during variant calling; use VCFtools instead for population genetics statistics."

cosmic-database

jaechang-hits

"Query COSMIC (Catalogue Of Somatic Mutations In Cancer) for cancer somatic mutations, gene census data, mutational signatures, drug resistance variants, and cancer gene annotations. REST API v3.1 supports gene/sample/variant queries. Free registration required. For germline clinical variants use clinvar-database; for drug-target data use opentargets-database or chembl-database-bioactivity."

contract-review-pro

CSlawyer1985

"专业合同审核 Skill，基于合同审核工作区成熟经验，提供7步工作流、终稿三件套、风险六维度评价"

accelint-ts-documentation

gohypergiant

Audit and improve JavaScript/TypeScript documentation including JSDoc comments (@param, @returns, @template, @example), comment markers (TODO, FIXME, HACK), and code comment quality. Use when asked to 'add JSDoc', 'document this function', 'audit documentation', 'fix comments', 'add TODO/FIXME markers', or 'improve code documentation'.

3d-cv-labeling-2026

curiositech

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

related-work-writing

lingzhi227

Write Related Work sections that compare and contrast prior work with your approach. Organize by theme, cite broadly, and explain how your work differs. Use when writing or improving the Related Work section of a paper.

addressing-pr-review-comments

streamlit

Address all valid review comments on a PR for the current branch in the streamlit/streamlit repo. Covers both inline review comments and general PR (issue) comments. Use when a PR has reviewer feedback to address, including code changes, style fixes, and documentation updates.

ghidra-rpc

cellebrite-labs

Reverse engineering assistant powered by Ghidra. Use for binary analysis, decompilation, vulnerability research, auditing compiled code, renaming symbols, annotating disassembly, cross-references, or any RE task - even without explicit mention of Ghidra.

Absurdist Voice

omer-metin

release-publish

mono

Publish SkiaSharp packages and finalize the release. Use when user says "publish X", "finalize X", "tag X", or "finish release X". This is the FINAL step - after release-testing passes. Publishes to NuGet.org, creates tag, GitHub release, and closes milestone. Triggers: "publish the release", "push to nuget", "create github release", "tag the release", "close the milestone", "annotate release notes", "testing passed what's next", "finalize 3.119.2", "release is ready".