skills

Plugin skills in Getty/skills. Commands read from the repo's manifest — yesterday.

Skills (115)

• ai-web-scraping

  Use when scraping, crawling or extracting structured data from websites, especially with an LLM in the loop — Crawl4AI, Firecrawl, ScrapeGraphAI, Scrapy, Crawlee, Scrapling, Playwright, Stagehand, browser-use, Jina Reader, Zyte; turning a site into a dataset or RAG corpus, monitoring pages for chang

• rocm-engineering

  Use when running, building or debugging GPU workloads on AMD hardware with ROCm — Instinct, Radeon or Ryzen APU graphics on Linux, Windows or WSL; amdgpu, TheRock, rocminfo, amd-smi, HIP, HIPIFY, porting CUDA code, PyTorch, JAX, ONNX, vLLM, SGLang, llama.cpp, Ollama, ComfyUI, training, quantization,

• vllm-operator

  Use when installing, running or tuning vLLM — small or consumer GPUs, DGX Spark, TTFT and throughput under concurrency, KV/prefix caching, quantization, OOM, tool calling.

• getty-agent-team

  Use when a repo needs subagents — setting one up, wiring briefing, adding karr, or when .claude/agents is missing or drifted from the house pattern.

• getty-skill-library

  Use when adding, naming, moving or renaming a skill in Getty's projects, installing one into a repo with skilletor, or before editing an installed SKILL.md.

• skill-authoring

  Use when writing or reworking a SKILL.md, or when a skill misfires — never triggers, fires on the wrong tasks, or is read but not followed.

• skill-compressor

  Use when a skill, prompt or instruction file must shrink — compressing one, merging overlapping ones, cutting token cost.

• skill-mining

  Use when a skill's content should come from existing code — mining conventions out of a codebase, or checking a skill against real practice.

• claude-code-plugin-engineering

  Use when building or debugging a Claude Code plugin — plugin.json, marketplace.json, hooks, mods, plugin MCP/LSP, userConfig — or when a plugin installs but does not load.

• claude-cross-session

  Use when one Claude session should reach another — claude --bg, claude agents, SendMessage/ListAgents, attach/logs/stop, or a background session that never answers.

• claude-headless

  Use when running Claude Code headless — claude -p, JSON output, claude calling claude, or a nested run that hangs, denies a tool or picks the wrong model.

• model-routing

  Use when choosing which Claude model runs work — opus/sonnet/haiku/fable for a subagent, headless run or Workflow stage, or setting an agent's model field.

• codex-cross-session

  Use when reaching an existing Codex session — codex queue, codex agents, remote-control start/pair, --remote, or a queued message that never arrives.

• codex-headless

  Use when running Codex non-interactively — codex exec, JSONL events, resuming a thread, sandbox settings, or an exec run that hangs on stdin or refuses a model.

• codex-plugin-engineering

  Use when building or debugging an OpenAI Codex plugin — plugin.json, .codex-plugin/, marketplace.json, hooks.json, plugin MCP — or choosing between plugin, Codex SDK and app-server.

• eu-ai-act

  Use when classifying an AI feature, system or model under the EU AI Act — provider or deployer role, prohibited practice, high-risk (Annex I/III), Article 50 transparency, marking AI-generated content, GPAI model duties, fine-tuning, AI literacy, conformity, AI Act deadlines.

• eu-ai-data-ip

  Use when onboarding a dataset or model, scraping or crawling for AI, fine-tuning, building RAG over third-party content, reviewing a model or software licence, deciding who owns AI output, or handling a copyright or rights-holder complaint — TDM exception, UrhG, trade secrets.

• eu-platform-accessibility

  Use when building or changing a consumer-facing UI, checkout, AI chat or agent interface, mobile app or generated document, or writing an accessibility statement or exemption — European Accessibility Act, BFSG/BFSGV, EN 301 549, WCAG, screen reader, keyboard.

• eu-platform-agent-assurance

  Use when reviewing a PR, release or agent change for compliance evidence on a regulated platform — evidence gates in CI, control tests, release go/no-go, agent tool permissions, feature flags, model or vendor swaps, runtime drift findings, check_evidence.py.

• eu-platform-billing-tax

  Use when building usage metering, token or subscription billing, invoices and credit notes, VAT, OSS or reverse charge, German e-invoicing (XRechnung, ZUGFeRD), accounting exports or retention (GoBD), PSP integration, prepaid credits, wallets or marketplace payouts.

• eu-platform-business-contracts

  Use when reviewing B2B contracts of an AI or SaaS platform — enterprise terms, DPA/AVV, SLA, security schedule, customer questionnaire or audit, model-provider or vendor terms, acceptable use, subprocessors, Data Act switching and exit, banks, health or public-sector customers.

• eu-platform-compliance

  Use when scoping or coordinating compliance for an EU AI or SaaS platform — launch readiness, market or B2C/B2B expansion, a new vendor or model, a regulated use case, or a change touching several legal domains (GDPR, AI Act, DSA, NIS2, consumer law, tax) where it is unclear which applies.

• eu-platform-consumer-commerce

  Use when building B2C signup, checkout, pricing, subscriptions, trials, renewals, withdrawal (Widerruf), the cancellation button (Kündigungsbutton) or consumer terms, or when a model or quota change affects paying consumers — EU consumer law, BGB §§ 312 ff. and 327 ff., PAngV.

• eu-platform-content-safety

  Use when users can publish, share or report content, a marketplace or agent store is added, minors may use the service, or a takedown notice, appeal or government or e-evidence order arrives — DSA, notice and action, statement of reasons, DDG, JMStV, JuSchG, age assurance.

• eu-platform-cybersecurity

  Use when designing or reviewing security of an EU AI platform — tenant isolation, agent tools, RAG, supply chain, vulnerability handling, shipping a desktop app, client, agent or self-hosted edition (CRA), NIS2/BSIG scope, or a security incident with a reporting deadline.

• eu-platform-data-governance

  Use when defining data retention, deletion or backups, tracing RAG, vector or training data, anonymising, onboarding a processor or subprocessor, or moving personal data outside the EU — SCCs, transfer impact assessment, adequacy, EU-US Data Privacy Framework.

• eu-platform-privacy

  Use when a feature processes personal data in the EU — controller or processor role, lawful basis, consent or legitimate interest, RoPA, DPIA, DPO, privacy notice, access or erasure request, profiling and automated decisions, a personal-data breach. GDPR, BDSG.

• eu-platform-regulatory-watch

  Use when checking whether EU or national law affecting an AI platform has changed — a new amendment, proposal or Omnibus, regulator guidance, a standard's citation, an applicability date, changed vendor terms, refreshing a legal source register, review_sources.py.

• eu-platform-tracking-marketing

  Use when adding or auditing cookies, local storage, analytics, pixels, SDKs, session replay, fingerprinting or server-side tagging, building a consent banner or CMP, or sending newsletters, email, SMS or phone marketing — ePrivacy, TDDDG, UWG, GDPR.

• fibo-integration

  Use when applying FIBO, the Financial Industry Business Ontology — financial concept definitions, modules and IRIs, OWL imports and releases, legal entities, instruments, listings, mappings, extensions.

• financial-identifiers

  Use when selecting, validating, looking up or reconciling financial identifiers — LEI and GLEIF, ISIN, MIC, CFI, FISN, tickers, entity vs instrument vs listing, checksums, lifecycle changes, licensing.

• geospatial-standards

  Use when exchanging, querying or validating geodata — coordinate reference systems, axis order, WKT and WKB, OGC Simple Features, GeoJSON, GeoPackage, GeoParquet, GeoSPARQL, OGC API Features, containment and distance queries.

• gs1-epcis

  Use when modeling, exchanging or validating GS1 EPCIS 2.x traceability events — CBV business steps and dispositions, GTIN, SSCC, GLN, Digital Link, aggregation, transformation, corrections, sensor data, capture and query interfaces.

• iptc-news-standards

  Use when creating, converting, ingesting or validating IPTC news data — NewsML-G2, ninjs, NewsCodes, Media Topics, RightsML, item revisions, corrections, cancellation, embargo, taxonomy mapping, claim extraction.

• openstreetmap-data

  Use when acquiring, filtering or joining OpenStreetMap data — nodes, ways, relations, tags, PBF and XML extracts, OsmChange diffs, geometry assembly, Nominatim geocoding, routing, tiles, ODbL obligations.

• wikibase-data

  Use when reading, exporting or ingesting Wikidata or Wikibase data — entity JSON, API, dumps, RDF, statements, ranks, snaks, qualifiers, references, truthy triples, redirects, deletions, update streams, reconciliation.

• xbrl-reporting-standards

  Use when reading, producing, converting or validating XBRL reports and taxonomies — XBRL 2.1, Dimensions, Inline XBRL, OIM, xBRL-JSON, xBRL-CSV, facts, contexts, units, DTS, SEC and ESEF filings.

• postgres-vector-graph

  Use when working with pgvector or Apache AGE on PostgreSQL — HNSW/IVFFlat, filtered or hybrid vector search, Cypher and agtype, or combining relational, vector and graph.

• feedback-loop-debugging

  Use when a bug, test failure or performance regression needs diagnosing — something broken, throwing, failing or suddenly slow.

• getty-create-software

  Use when creating, scaffolding or initializing a new project or module — Perl dist, Node app, Python package, Go binary, or a bare .gitignore.

• ai-playable-games

  Use when building or reviewing a game, simulation or prototype that scripted bots, search AIs or LLM agents must play — the observation and order interface, a bot/agent API or MCP server for a game, AI seats next to human players, fog of war for agents, replaying agent games; or when an AI player ch

• game-balance-math

  Use when pricing or tuning game objects and systems by the numbers — cost curves for units, ships, buildings or upgrades; counter tables and payoff matrices; dominant or useless strategies; designing or auditing an economy's sources, sinks, converters and trade; inflation, hoarding, infinite-money l

• simulation-balancing

  Use when balancing a game through bot, AI-vs-AI or self-play runs instead of (or before) human playtests — batch runners, seed sets, bot policy leagues, parameter sweeps, win-rate comparisons; when asking whether a mechanic matters or can be cut; when a rule change shows no effect in the sim; or whe

• getty-git-commit-style

  Use when writing or amending a commit message or a Changes/CHANGELOG entry in a Getty repository, stamping a release's version line, or committing across several repos.

• getty-git-usage

  Use when committing, pushing or landing work in a Getty repo — rebase vs merge, history cleanup, stacked branches, releases. Commit wording: getty-git-commit-style.

• hermes-plugin-engineering

  Use when extending NousResearch Hermes Agent — a plugin, tool, hook, middleware, provider, gateway or Telegram adapter, or a Desktop/Dashboard extension.

• electron-desktop-engineering

  Use when building, securing, debugging, packaging or updating an Electron app — main, preload and renderer, contextBridge, IPC, sandbox, CSP, custom protocols, native modules, Electron Forge, electron-builder, electron-vite, fuses, code signing and notarization, auto-update.

• react-ecosystem-engineering

  Use when building, reviewing, debugging, migrating or testing a React app — hooks, React Compiler, SSR/RSC, Next.js and other frameworks, routing, state, data fetching, forms, UI libraries, choosing dependencies, React in an Electron renderer, or React Flow (nodes, edges, handles, layout).

• getty-perl-core

  Use when editing any Perl in a Getty project — module loading, subs, errors, style, cpanfile pins, per-file $VERSION.

• getty-perl-distribution

  Use when creating a Perl CPAN distribution or bringing an existing one in line with [@Author::GETTY] — dist.ini, cpanfile, Changes, lib/, t/, CI scaffolding.

• getty-perl-moo

  Use when writing classes or roles in a Moo distribution — `use Moo`, `use Moo::Role`, attributes, roles, delegation, lifecycle.

• getty-perl-moose

  Use when writing classes or roles in a Moose distribution — `use Moose`, `use Moose::Role`, attributes, modifiers, immutability.

• getty-perl-pod

  Use when writing or editing POD in a distribution whose dist.ini uses [@Author::GETTY] — the =attr/=method/=opt shortcut commands, the # ABSTRACT line, and which sections Pod::Weaver generates so you must not write them.

• getty-perl-release-author-getty

  Load when a dist.ini contains [@Author::GETTY] — bundle options, POD conventions (=attr/=method/=opt), next-version semantics, the dzil release workflow.

• getty-perl-sugar

  Use when designing a new Perl DSL or sugar — Moose::Exporter keywords, declarative structures, metadata registries. Not for using an existing one.

• getty-perl-typing

  Use when typing Perl attributes or parameters — isa, Type::Tiny, Types::Standard, Moose constraints, or whether a project needs types at all.

• perl-alien

  Use when providing a C library or tool through Alien::Build — an alienfile, system vs share builds, pkg-config probing, or consuming Alien::Base from XS or FFI.

• perl-io-async-future

  Use when writing async Perl — IO::Async, Future, Future::AsyncAwait, Net::Async::*, futures lost to GC, cancellation, requests that hang.

• perl-mcp

  Use when building an MCP server in Perl — MCP::Server, tool definitions, server setup, or wiring MCP into a Perl LLM client such as Langertha.

• perl-mojo

  Use when writing Mojolicious or Mojo::Base code — attributes and defaults, roles, Mojo::UserAgent, Mojo::IOLoop, the core Mojo::* helpers.

• perl-par-binary-release

  Use when a Perl dist should ship standalone PAR::Packer (pp) binaries in its GitHub release: release-binaries workflow, build/verify/smoke scripts, runtime-lib check.

• perl-release-dist-ini

  Load when reading, editing or debugging any dist.ini — plugins, bundles such as Author::GETTY, version config, prereqs.

• perl-xs

  Use when writing or debugging XS — .xs files, typemaps, ppport.h, xsubpp or C compile errors, segfaults or leaks at the Perl/C boundary.

• jsonld-processing

  Use when writing or debugging JSON-LD 1.1 — @context, term expansion, @id and @type, language and datatype coercion, @list and @set, named graphs, expand, compact, flatten, frame, RDF conversion, remote contexts.

• owl2-engineering

  Use when designing, inspecting or reasoning over an OWL 2 ontology — classes, properties, axioms, restrictions, cardinality, open-world semantics, owl:sameAs, DL and EL/QL/RL profiles, imports, reasoners.

• prov-o-provenance

  Use when modeling, recording or querying provenance with W3C PROV and PROV-O — entities, activities, agents, wasDerivedFrom, wasAttributedTo, qualified relations, revisions, bundles, provenance of provenance.

• rdf-linked-data

  Use when modeling, parsing, publishing or debugging RDF — triples, IRIs, blank nodes, literals and datatypes, RDFS, Turtle, TriG, N-Triples, N-Quads, named graphs, Linked Data, RDF 1.1 vs 1.2.

• shacl-standards

  Use when writing, running or debugging SHACL validation — shapes, targets, property paths, Core constraints, sh:closed, qualified value shapes, SHACL-SPARQL, validation reports, SHACL 1.2.

• skos-vocabularies

  Use when building, mapping or validating a SKOS vocabulary — thesaurus, taxonomy, concept scheme, multilingual labels, notations, broader/narrower, exactMatch and closeMatch, collections, SKOS-XL, spreadsheet migration.

• sparql-querying

  Use when writing, optimizing or debugging SPARQL queries and updates — SELECT, ASK, CONSTRUCT, OPTIONAL, FILTER, unbound variables, aggregates, property paths, GRAPH, SERVICE federation, pagination, SPARQL 1.2.

• fritzbox-operator

  Use when integrating, automating or debugging a FRITZ!Box or FRITZ! smart-home devices — TR-064, AHA HTTP, the Smart Home REST API, login or SID errors, Home Assistant FRITZ!Box Tools or FRITZ!SmartHome, DECT thermostats, Control 440 buttons, presence trackers, mesh, guest network, provider-managed

• home-assistant-automations

  Use when writing or debugging a Home Assistant automation, script, scene, helper, blueprint, template or dashboard — triggers firing twice or not at all, delays lost on restart, manual override, lighting, heating, leak response, presence, energy, Assist or LLM/MCP control, traces.

• home-assistant-hacs

  Use when installing, updating, removing or debugging HACS or a Home Assistant custom integration, card or theme — custom repositories, a card that fails to load (\"Custom element doesn't exist\"), frontend resources, Mushroom, Bubble Card, card-mod, Adaptive Lighting, LocalTuya, Tuya Local, SonoffLA

• home-assistant-operator

  Use when installing, upgrading, backing up, migrating or debugging Home Assistant itself — HAOS, Container, Supervisor, apps and add-ons, restore, device and entity registries, REST or WebSocket API and tokens, recorder database growth, discovery, slow startup or high CPU, custom integration develop

• smart-home-architecture

  Use when designing, expanding or migrating a smart home around Home Assistant — choosing protocols or hubs, local versus cloud control, deciding which smart-home skill applies, buying devices from Amazon or EU listings, comparing compatibility, resilience and failure testing, VLANs and discovery.

• smart-home-devices

  Use when identifying, choosing, integrating or debugging a specific smart-home product in Home Assistant — Philips Hue, Shelly, THIRDREALITY, Tuya and _TZ Zigbee fingerprints, SONOFF/eWeLink, Zengge, ThermoBeacon, Aqara/Xiaomi, OSRAM/LEDVANCE, Govee, Paulmann, thermostats, plugs, locks, covers, pres

• smart-home-esphome-mqtt

  Use when writing ESPHome YAML, flashing or recovering an ESP32, ESP8266 or LibreTiny device, OTA or safe mode, Bluetooth proxies, or designing MQTT topics, Mosquitto, discovery, retained messages and availability for Home Assistant.

• smart-home-radio

  Use when a Zigbee, Thread, Matter, Bluetooth or Z-Wave network misbehaves or changes — pairing fails, devices drop off, ZHA or Zigbee2MQTT, coordinator migration, routers and binding, Zigbee OTA, Matter commissioning, Thread border routers, IPv6, BLE proxies, EnOcean.

• smart-home-video-media

  Use when integrating or debugging cameras, NVRs, doorbells, chimes, TVs or media players in Home Assistant — Reolink, RTSP, ONVIF, WebRTC, go2rtc, Frigate, snapshots and notifications, LG webOS, Fire TV, Android TV, Sonos, Music Assistant.

• linkedin

  Use when working on anything LinkedIn — posts, content strategy, profile or company page, algorithm and limits, ads, recruiting, analytics, or the DACH legal requirements.

• twitch
• storytold-artcraft
• storytold-artcraftx
• storytold-cadcraft
• storytold-craft-launcher
• storytold-craft
• storytold-deckcraft
• storytold-designcraft
• storytold-effectcraft
• storytold-filmcraft
• storytold-gridcraft
• storytold-lightcraft
• storytold-pdfcraft
• storytold-photocraft
• storytold-soundcraft
• storytold-vectorcraft
• storytold-wordcraft
• cisco-ios-router
• cisco-small-business-switch
• docker-engine-api
• docker-registry
• docker
• gearman
• hetzner-operator
• kubernetes-cilium-concepts
• kubernetes-concepts
• kubernetes-gpu
• kubernetes-rke2
• mikrotik-routeros
• network-device-ssh-cli
• proxmox-ve
• rex
• samba-active-directory
• ssl-tls-pki
• storage-engineering