astra-technology-company-limited
33 verified skills
autorun
ASTRA mostly-autonomous execution — runs the entire pipeline from planning through PR merge and worktree removal with minimal user input, auto-iterating up to N times until tests pass. Sequentially executes /service-planner (with HTML mockup screens) → /blueprint (which in v5.10+ runs in a worktree-first order: creates the sprint worktree via /sprint-init --scaffold-only, then authors + reviews + commits the blueprint inside the sprint worktree) → /sprint-init (idempotent re-entry — usually a no-op in the standard path) → /test-scenario → implementation (/generate-entity + blueprint-based) → /test-run → /pr-merge --auto → automatic worktree removal; on test failure it classifies the cause and re-enters from the appropriate stage (self-improvement loop). Every user-choice step is auto-decided via smart defaults except for two HITL points: max-iteration count (asked once at start) and the post-merge promotion target (dev/staging/skip — asked at the end of Stage 8 before final cleanup, because the deployment surface choice has no safe default). Other HITL fires only on true blockers — missing gh authentication, merge conflicts, or Critical review issues. Use when you want a single command to drive a week's worth of work with one pause for the promotion-target decision.
development
sprint-init
Initializes a new ASTRA sprint. Creates an isolated sprint worktree (with port-isolated dev server settings), generates sprint prompt maps, progress trackers, and retrospective templates inside that worktree, and prints the cd path so all subsequent development and testing happens in the worktree. With --auto flag, also auto-executes the post-scaffolding pipeline: /test-scenario → implementation → /test-run → /pr-merge --auto (worktree auto-removed). Between each major stage (5.2/5.3/5.4/5.5 iteration/5.6), the skill performs a silent save (auto-state.yaml + commit) and applies a 'reference-avoidance' rule (don't re-read large prior artifacts; rely on yaml SSoT) so the system's built-in auto-compression keeps context manageable, then continues directly to the next stage without user intervention. --resume flag is reserved for true recovery (context crash, forced interrupt) — it reads auto-state.yaml and jumps to next_stage. Only halts on true blockers (gh auth, merge conflicts, Critical review issues).
development
tool-author
Authors and validates LLM tool descriptions and input schemas (Anthropic Tool Use, MCP servers, LangChain @tool, Pydantic, Zod). Use when the user mentions "tool description", "function calling", "MCP tool", "Pydantic schema", "Zod schema", "@tool decorator", "input_schema", "tool spec", "툴 정의", "함수 호출 스키마", or when editing files that define LLM tool surfaces. Enforces the six required attributes (one-line summary, anti-pattern, synonyms, parameter examples, enum constraints, return shape) and blocks the seven known failure modes — wrong-tool selection, skipped tool, malformed arguments, retry loops, user-intent bypass, wrong side-effect, and un-auditable traces. For authoring ASTRA SKILL.md files use /skill-author instead — this skill is for *runtime* LLM tool surfaces, not for skill files themselves.
tools
manual-generator
Automatically generates a professional online service manual or help center from a running service URL and project documents. Captures per-screen screenshots via Chrome MCP, extracts feature descriptions from blueprints / planner documents, and publishes step-by-step guides as a self-contained HTML package under docs/manual/{feature-name}/. Use when generating a manual, writing a user guide, producing help documentation, or building a help center landing page.
tools
user-test
Performs end-user UAT (User Acceptance Testing) by driving a real browser through Chrome MCP, self-verifying each step with hard assertions (DOM / Network / URL / Console), auto-assigning severity on failure, and emitting an HTML report plus issues.md into a timestamped session folder. Supports two modes: interactive (URL + Vietnamese natural-language flow description) and --auto (batch-run pre-authored test cases under docs/tests/uat-cases/). Use when the user asks for "UAT", "user acceptance test", "kiểm thử người dùng", "regression test", or runs /user-test, /uat. Distinct from /test-run (developer-authored technical integration testing) and /test-scenario (scenario authoring from blueprints).
tools
uat-parallel
Runs UAT (User Acceptance Testing) cases in TRUE PARALLEL using Playwright Test runner with isolated browser contexts per worker (separate cookies, localStorage, sessionStorage). Solves the two main limits of /user-test: (1) sequential single-page execution that does not scale beyond a few cases, and (2) one stuck case blocking the rest of the run. Reuses 100% of the /user-test UAT case Markdown+YAML format under docs/tests/uat-cases/, runs them via `npx playwright test --workers=N`, and emits the same report layout (index.html + issues.md + session.json + screenshots/) under docs/tests/uat-reports/. Use when the user asks to "run UAT in parallel", "speed up UAT", "test multi-user", "song song", "uat parallel", or runs /uat-parallel. Distinct from /user-test (sequential Chrome MCP, supports interactive mode), /test-run (developer integration tests), /test-scenario (scenario authoring).
tools
ai-agent-module
AI 에이전트 플랫폼 모듈을 자동으로 구축합니다. 레퍼런스 설계 문서를 기반으로 블루프린트 작성, 스프린트 생성, 구현, 테스트 시나리오 작성, 테스트 실행 및 디버깅까지 전체 파이프라인을 자동 실행합니다. 멀티 프로바이더 LLM 클라이언트, SSE 스트리밍, 도구/플러그인 시스템, 스킬 시스템, 에이전트 Core Loop(4중 안전장치), RAG/메모리, HITL, 서브에이전트, 모델 라우팅/폴백, 채널 연동 기능을 포함합니다.
data-ai
astra-setup
Sets up the ASTRA methodology global development environment (Step 0.0). Configures ~/.claude/settings.json, MCP servers, and required plugins.
tools
project-checklist
Verifies the ASTRA Sprint 0 completion checklist. Checks required files, settings, and quality gate configurations.
testing
auth-module
인증 모듈을 자동으로 구축합니다. 레퍼런스 설계 문서를 기반으로 블루프린트 작성, 스프린트 생성, 구현, 테스트 시나리오 작성, 테스트 실행 및 디버깅까지 전체 파이프라인을 자동 실행합니다. 회원가입, 로그인, 로그아웃, 토큰 관리, 약관 관리, 사용자 관리, 프로필 관리 기능을 포함합니다.
tools
code-standard
Automatically provides support for writing code based on ISO 3166-1/3166-2 and ITU-T E.164 standard data when implementing components that handle international standard codes such as phone number inputs, country selectors, and region selectors.
development
payment-module
구독 결제 모듈을 자동으로 구축합니다. 레퍼런스 설계 문서를 기반으로 블루프린트 작성, 스프린트 생성, 구현, 테스트 시나리오 작성, 테스트 실행 및 디버깅까지 전체 파이프라인을 자동 실행합니다. 플랜 관리, 결제 수단(빌링키), 구독 관리, 청구서/결제, PG사 연동(토스페이먼츠/KCP), 웹훅, 정기결제 스케줄러, Dunning(재시도), 크레딧 관리, 비례 배분(Proration) 기능을 포함합니다.
tools
slack-to-sprint
Slack 채널의 List 항목을 분석하여 블루프린트와 스프린트 프롬프트 맵을 자동 생성합니다. 채널 선택 → List 선택 → Item 선택 → 상태 업데이트 → 요구사항 분석 → 블루프린트/스프린트 생성의 워크플로우를 제공합니다.
business
sprint-plan
Initializes a new ASTRA sprint. Creates sprint prompt maps, progress trackers, and retrospective templates.
data-ai
ux-publish
기획 산출물(service-planner)의 IA/화면설계서와 기능정의서를 기반으로 프로덕션 수준의 UI 컴포넌트를 생성합니다. 프로젝트의 기술 스택(React/Vue/Angular/Svelte)에 맞는 컴포넌트를 publish/ 디렉토리에 생성하고, 이후 개발 시 각각의 실제 위치(src/)로 복사하여 사용할 수 있는 구조를 제공합니다. Vibe Coding 디자인 가이드와 애니메이션 가이드를 기반으로 세련된 프로덕션급 UI를 구현합니다.
development
workspace-module
워크스페이스 모듈을 자동으로 구축합니다. 레퍼런스 설계 문서를 기반으로 블루프린트 작성, 스프린트 생성, 구현, 테스트 시나리오 작성, 테스트 실행 및 디버깅까지 전체 파이프라인을 자동 실행합니다. 워크스페이스 CRUD, 멤버 관리, 초대 시스템(이메일/링크), 소유권 이전, 워크스페이스 전환, 인증/결제 모듈 연동 기능을 포함합니다.
tools
astra-guide
ASTRA methodology quick reference guide. Displays workflow, command, quality gate, handoff process, and behavioral guardrail summaries.
testing
coding-convention
Automatically applies the project's coding conventions when writing or modifying code. Used when editing or creating Java, TypeScript, React Native, Python, CSS/SCSS files, or when performing code reviews and refactoring.
development
catalog-generator
Automatically generates a professional product promotional catalog as a self-contained HTML package from product data. Executes the full catalog production pipeline — planning, data normalization, design, copywriting, and validation — in a single pass without user feedback.
development
design-extract
Extracts an OKLCH color palette, font families, and spacing heuristics from design references such as images, PDFs, URLs, and screenshots, and produces a draft DESIGN.md Front Matter report. Uses Vision MCP (fect-mcp__vision_analyze, vision_ocr) or WebFetch to analyze references. Inputs: local file paths, URLs, Figma screenshots, competitor sites, brand guideline PDFs. Output: docs/design-system/extract-report-{date}.md (a draft DESIGN.md Front Matter YAML + extraction evidence). This skill only generates the report — it never edits DESIGN.md directly. Merging is performed when the user, after reviewing the report, explicitly invokes /design-init --apply-extract=<report-path>. /design-init --from-refs=... also uses a one-way single-call flow where this skill is invoked once internally and the resulting report is merged.
tools
design-init
Creates/updates docs/design-system/DESIGN.md — the single source of truth (SSoT) for the project's design system — and auto-generates src/styles/design-tokens.css. Fills OKLCH color, typography, spacing, and motion tokens from brand keywords, persona, and reference inputs (images/URLs), and codifies the anti-AI aesthetic rules (Vibe Coding). Outputs a hybrid format (YAML Front Matter + Markdown Body) that combines the Google Stitch DESIGN.md spec with the ASTRA 3-tier token structure (Primitive→Semantic→Component). Modes: new generation (init), update on existing (update), regenerate CSS only (--regenerate-css), extract from references then merge (--from-refs=<path-or-url> — internally calls /design-extract once), and merge from a pre-existing extract report (--apply-extract=<report-path>). If docs/planner/*/interview-report.md is present the persona is auto-cited, and /service-planner, /handoff-publish, and design-token-validator all reference this file as the SSoT.
development
pr-merge
Runs an automated iterative cycle from PR creation through code review, issue fixes, and merge. v5.11+ uses an integration-branch model: sprint PRs (head=feat/sprint-*) target a feature- or fix-level integration branch (feat/<name> or fix/<name>) chosen interactively (pick from existing OR create new with user-chosen base). After the sprint PR merges into the integration branch, the user always picks a promotion path — staging (fast/hotfix) or dev (standard) or skip — and a promotion PR is created and merged without a second review. In a sprint worktree, this command handles commit → push → PR-create → code-review → fix → re-review and then stops; the merge runs from the main worktree. With --auto, the command transitions phases by itself and uses safe defaults (reuse existing integration branch if name matches, else create new from dev), but the final promotion-target branch is always HITL even under --auto. Confirmation prompts auto-approve except for safe HITL points (gh authentication, merge conflicts, Critical issues, ambiguous multi-PR detection, promotion target).
development
skill-author
Creates new SKILL.md files or refactors existing skills to comply with the ASTRA skill best practices guide (docs/development/skill-best-practices.md). Use when user mentions "new skill", "create skill", "SKILL.md", "skill authoring", "스킬 작성", "스킬 만들기", or when editing any file matching skills/**/SKILL.md.
development
project-init
ASTRA Sprint 0 project initial setup. Supports Web and Mobile (React Native, Flutter, KMP) platforms. Creates project directory structure, CLAUDE.md, design system templates, blueprint templates, and sprint templates.
development
sprint-progress
Automatically updates sprint progress tracking when blueprint, database design, test case, implementation, or test report files are created or modified. Used when writing files under docs/blueprints/, docs/database/, docs/tests/, docs/sprints/, or src/ directories.
testing
data-standard
Automatically applies the public data standard terminology dictionary when writing database tables, entity classes, SQL, and migration files. Uses standard English abbreviations for DB column names, table names, and variable names, and complies with domain rules. Used when writing ORM entities, DTOs, VOs, SQL DDL, and schema definitions.
data-ai
slack-import
Analyzes Slack channel List items to auto-generate blueprints and sprint prompt maps. Provides a workflow of channel selection → List selection → Item selection → status update → requirements analysis → blueprint / sprint generation.
data-ai
service-planner
Auto-generates planning deliverables for a feature. Runs the full planning pipeline based on the Design Thinking methodology: market analysis → actor derivation → persona interviews → pain-point analysis → idea derivation (HMW/SCAMPER/JTBD) → requirements definition (KPI/OKR) → use-case definition (journey maps) → IA / screen design → design-system-applied HTML mockup screens → feature definition (story map / risks).
development
design-redesign
Audits existing UI components/pages/CSS against docs/design-system/DESIGN.md and applies fixes to restore design consistency. Uses design-token-validator to detect hardcoded color/font/spacing violations, and invokes designer-persona to report a senior-perspective score (0-10) and anti-AI aesthetic violations (generic shadcn look, purple gradient cliché). Fixes are applied automatically (--apply) or proposed as a PR (--pr); after changes, design-token-validator is re-run and must PASS before completion. Input: target directory (e.g., src/components/Button), single file, or git diff. Output: audit report (docs/design-system/audit-{date}.md) + fix proposals + application results.
tools
handoff-publish
Generates a handoff package (`{feature}-handoff/` at the branch root) for UX/UI/Dev/QA collaboration. Identifies screens via the 4-segment Screen ID scheme (DOMAIN-PAGE-SECTION-UC) and composes 11 documents centered on the SSoT `1-screen-registry.md` (state matrix, edge cases, responsive, component specs, business rules, UX writing, IA/sitemap, personas, Decision log). If service-planner deliverables exist, they are automatically converted and reused.
development
test-run
Launches the server and performs integration testing with a real browser. Supports cmux built-in browser (primary) and Chrome MCP (fallback). Automatically conducts server log monitoring, page verification, API behavior checks, and performance measurement.
tools
test-scenario
Creates E2E test scenarios based on the project's existing work context. Analyzes blueprints, DB design, routes, and API endpoints to generate comprehensive test scenarios for integration testing.
development
blueprint
Authors a Blueprint (design document) for a feature. v5.10+ runs in a worktree-first order: (1) determine slug + blueprint directory number on the main worktree, (2) auto-create the sprint worktree by delegating to /sprint-init --scaffold-only, (3) cd into that worktree, (4) author the blueprint inside the worktree (10 standard sections — data flow, schema, API contract, sequence, pseudocode logic, HITL Triggers — implementation code excluded), (5) blueprint-reviewer verifies quality, (6) commit to the sprint branch. If /service-planner deliverables (docs/planner/{NNN}-{slug}/) exist, they are auto-loaded. Only 1–3 core decisions that genuinely require human judgment (PK strategy, transaction boundary, sync/async for external dependencies) are asked via AskUserQuestion. Section 10 (HITL Triggers) is consulted by /feature-dev during implementation so the user is only asked on essential decisions. Already inside a sprint worktree → worktree creation skipped (secondary blueprint case). On non-standard branch (not dev/main/master) → error with checkout guidance.
development