Coding & Development

Works a local issue board shared by a team of coding agents through the earmark CLI. Joins the board as a worker, reviewer or pm profile, claims the next issue in scope, reads only what changed, hands work off to review, files work only another specialist can do (images, audio, copy) as blocking issues, approves or rejects handoffs, and triages and assigns the backlog. Use when the project has a .earmark/ board, when the user mentions issues, tickets, a backlog, the board, a review queue or earm

Builds and maintains PHP APIs with Kipchak, the API development kit on Slim 4 and FrankenPHP worker mode. Covers routes, controllers, DTOs with OpenAPI attributes, models, tests, configuration, drivers and middleware, and writing new drivers and middleware. Use when a project's composer.json requires kipchak/core, when the user mentions Kipchak, or when asked to build a lean, explicit PHP API with Kipchak.

Portable outer-operator procedure for durable Argus missions across five hosts.

Open a local page with the context tree of this project, for each mode of the main session and each agent, showing what is loaded at start (Claude Code's own files, the hecpack hook, the agent body) and what it may read by link. It also lists the files that cost most, the rules by path, the docs nobody reads and the repeated text, with the change since the last run. A script builds it with no model call. Use when the user runs /hecpack:context or asks how much context an agent or mode starts wit

Coordinate coding agents on a shared local project with task ownership, file claims, Git checkpoints, context handoffs, independent review and recovery. Use when the user requests collaboration between agents such as Codex, Claude Code or Antigravity.

Writes user stories as Markdown files in the standard format and uploads them to Taiga, each with the tasks listed under "## Tasks". Use when the user asks to write, create, standardize or review a story .md, or to upload, register or push stories and tasks to Taiga, a sprint or the backlog.

Gouverne, conçoit, crée et raffine les skills, les règles (.agents/rules/) et les standards normatifs selon les patterns officiels d'Anthropic, et maintient la conformité SemVer.

Cadre, structure et valide la réalisation de Labs techniques pour les prototypes de projets fil rouge, propose les skills IA à créer avec plugin-dev, génère le plan idéal de présentation (avec page de garde et 3 phrases de notes orales A1 par slide) et le fichier Todo.md, sans rédiger la solution technique à la place de l'apprenant.

Guided learning mode. The user names something they want to do or build, and instead of doing it for them, guide them step by step with micro challenges, questions, hints and curated learning resources, so they understand and own the result. Invoke only when the user types /know-nothing, or explicitly asks to be guided through a task so they learn it rather than having it done for them.

Set up an Australian proprietary company (Pty Ltd) from scratch so it gets an ACN and ABN. Use whenever the user wants to incorporate or register a company in Australia, get an ACN or ABN for a new company, draft a company constitution, ASIC Form 201 data sheet, consents to act as director, application for shares, share certificates or statutory registers, or plan the tasks to incorporate. It runs an onboarding questionnaire, generates placeholder-safe document drafts, saves them to Google Drive

UX knowledge for building and reviewing user interfaces in React + Mantine web apps, distilled from GitHub's Primer design guidance. Use it whenever a task adds or changes something a user sees or does: a page, dialog, form, table, navigation, button, message, or the empty, loading and error states around them; when planning a feature's flow; and when reviewing UI in a pull request. It tells you which Mantine component fits which situation, which states a screen needs, how to lay out and word it

Find grocery products and fill the user's K‑Ruoka shopping list through the site's browser interface. Use when the user asks to add products for a K-Market, K-Supermarket, or K-Citymarket store selected in K‑Ruoka, using an Agent Skills-compatible AI tool; not for ordering or purchasing products.

Start a new iOS app, or choose and add iOS open-source libraries, using the awesome-ios-for-agents catalog. Asks the user what the app needs, picks Swift packages by objective data (activity status, stars, minimum iOS, Swift tools version), creates the project with a verified recipe (XcodeGen, Tuist, MobileBuildMCP or Xcode MCP), then builds and launches it in the iOS Simulator. Use when the user wants to create an iOS app, set up an Xcode project from scratch, or asks which library to use for n

База знаний по книге «Цель. Процесс непрерывного совершенствования» (The Goal) Элияху М. Голдратта и Джеффа Кокса: теория ограничений (TOC), цель и три показателя (пропускная способность, запасы, операционные расходы), зависимые события и статистические колебания, узкие места, запуск работы по ритму ограничения, буферы, пять фокусирующих шагов, политические ограничения, сократовский метод. Использовать, когда нужно найти ограничение (узкое место) в любом процессе; когда бизнес или команда не рас

Use when drafting or revising the title and description of a merge request or pull request, on any hosting service (GitLab, GitHub or another, cloud or self-hosted), before creating or updating it.

База знаний по книге «Атомные привычки» (Atomic Habits: Tiny Changes, Remarkable Results) Джеймса Клира. Использовать, когда нужно применить методики Клира: как выработать привычку и не бросить, как держать регулярность (каждый день писать, заниматься, учиться), как избавиться от вредной привычки, системы вместо целей, привычки на основе идентичности, четыре закона изменения поведения, среда и триггеры, наслоение привычек, правило двух минут, трекинг привычек, что делать, когда человек постоянно

Use when orchestrating Headless CMS content via MCP tools.

Make a small open LLM (e.g. Qwen3.5-0.8B) better at a goal by editing its weights directly, with no training ("weight-space programming"). Ready presets for maths (MATH-500) and science (MMLU-Pro); any other benchmark or behaviour works from your own prompts and matched example pairs (format, language, verbosity, sycophancy, loops). Finds the direction that separates the unwanted behaviour from the wanted one, edits the MLP down_proj columns that write it in a copy of the model on disk, screens

/sigma-implement <issue-number> builds one ticket on one branch and one worktree, then pushes. This skill does not review. Review is /sigma-review in another chat.

/sigma-create-ticket files or edits one SPEC, EPIC, or TICKET issue. It is also the gateway for /sigma-to-spec and /sigma-to-tickets: those skills hand it a body. /sigma-unslop does not. This skill is the only one that runs gh issue create or gh issue edit. Use when the user invokes /sigma-create-ticket or one of those skills hands a body.

Crea, revisa y explica scripts de k6 (Grafana k6) para pruebas de performance siguiendo el estándar del equipo — load test, stress test, smoke, spike, soak y breakpoint, tanto HTTP contra APIs como con k6 browser (Chromium, Web Vitals) o híbridos. Genera la carpeta k6/ con scripts de nombres estandarizados, flujos reutilizables, thresholds, checks, manejo seguro de credenciales y un .md explicativo por script pensado para gente que no es experta en k6. Usá esta skill siempre que el usuario menci

Use a short waiting gap while an AI agent works for one useful human step in current work, an unfinished task, or relevant learning, with a checkpoint and an easy return. Use for “what can I do while this runs?” or “I have ten minutes while my agent works”; not general day planning, agent monitoring, or end-of-day reflection.

Building Figma and FigJam content from a plugin — the sizing and font rules that fail silently, the clamps that stop one bad value killing a build, reading comments back through the REST API, and the loop that tells you whether your code actually ran. Use when writing or debugging a Figma plugin, generating frames or wireframes programmatically, or when a build aborts mid-draw, a layout collapses, text wraps one character per line, frames overlap something already on the page, or it "looks the s

Check a repository before it goes open source and fill in what is missing - license, README, SECURITY.md, CONTRIBUTING, issue and PR templates, package and Agent Skill metadata, CI and workflow safety, secrets and personal details in files and in the whole git history, and GitHub settings such as private vulnerability reporting and branch protection. Use it whenever someone is about to make a repository public, publish a first release, asks "is this ready to open source", "开源前检查", "准备开源", "还缺什么"