hashtagf-skills

Plugin hashtagf-skills from marketplace hashtagf in hashtagf/skills. Commands read from the repo's manifest — 5 days ago.

Skills (6)

• design-system-builder

  Build, extract, repair, or extend a complete design system — design tokens, a component library with full interaction states, and docs — delivered as CSS/Tailwind or as an installable package for React, Next.js, or Vue (auto-detected from the repo) — including publishing and sharing it as an npm pac

• figma-cli

  Build, edit, render, and verify Figma designs through figma-ds-cli's JSX DSL and Figma Desktop bridge. Use this skill whenever a task involves a figma-ds-cli project, its Frame/Text/Icon/Rectangle/Image/Instance/Slot JSX, `node ~/figma-cli/src/index.js` commands, `var:` token bindings, DESIGN.md imp

• gitops-nova

  SOP for designing, bootstrapping, extending, and auditing an Argo CD GitOps repository — app-of-apps layout, AppProjects as blast-radius boundaries, ONE shared service Helm chart driven by per-(service, env) values, ApplicationSet generators, the three-tier sync ladder (auto / manual / observe-only)

• kimi-k3

  Send a prompt to Moonshot's Kimi K3 model and relay its answer, using the bundled runner script. Use this skill whenever the user wants to ask, query, run, or delegate something to Kimi K3 / Kimi / the Moonshot model — e.g. "ask kimi-k3 to…", "get a second opinion from kimi", "run this through kimi

• test-design

  SOP for designing and running software tests at the right depth — test scenarios, test case checklists, test plans, coverage audits, bug reports, UAT scripts, and release go/no-go with production smoke and rollback — via a 4-level system (L1 smoke → L4 critical) that adapts to the kind of work (new

• ux-ui-principles

  Load before any UX/UI answer, even a one-line fact (WCAG target size, contrast) or a question about the user's own component, page or code (states, layout, copy). Holds rules general knowledge gets wrong: WCAG exceptions, Thai typography, money and approval flows, validation, component choice, requi

Subagents (2)

• kimi-k3-implement

  Hand a scoped coding task to Kimi K3 running as an autonomous agent (its own read/edit/run loop via Kimi Code CLI, model kimi-k3), then verify K3's work and report. Use this when the user wants Kimi K3 to actually BUILD/implement something — "let k3 implement this", "ให้ k3 ทำ feature นี้", "delegat

• kimi-k3

  Delegate a task to Moonshot's Kimi K3 model and relay its answer, by running the bundled runner script. Use this agent whenever the user wants Kimi K3 / Kimi / the Moonshot model to handle a prompt — "ask kimi-k3 to…", "ให้ k3 ช่วยดู…", "get kimi's take", a second opinion, a cross-model check, a lon

Similar plugins