원클릭으로
ultra-dev
ultra-dev에는 Petrina99에서 수집한 skills 11개가 있으며, 저장소 수준 직업 범위와 사이트 내 skill 상세 페이지를 제공합니다.
이 저장소의 skills
Generates professional end-user or developer project documentation as a PDF with table of contents, working internal/external links, and image placeholders. Does NOT auto-trigger and is NOT chained from other skills — runs only when the user invokes `/project-docs` explicitly. Scans the repo (routes, entry points, public APIs, CLI commands, config) and writes Markdown + a rendered PDF into `docs/user-guide/` or `docs/dev-guide/`. Image folder is auto-gitignored.
Build or update a branded end-user PDF manual under docs/user-manual/<slug>/ — prose + real annotated screenshots (Playwright) of the running app + brand/logo + optional legal notice + optional AES-256 protected PDF. Slash-only, never auto-triggers. Standalone from the core ultra-dev chain.
Execute a written `docs/ultra-dev/<slug>/plan.md` end to end with configurable branch, worktree, subagent dispatch, and commit-granularity options. Does NOT auto-trigger; runs only when chained from `spec-to-plan` or invoked explicitly via the Skill tool by name.
Reads an existing spec.md and produces docs/ultra-dev/<slug>/plan.md with Tasks, Dependencies, and Verification sections. Does NOT auto-trigger; runs only when chained from spec-writing or invoked explicitly via the Skill tool by name.
Generate documentation artifacts (changelog, README, inline code docs, ultra-dev notes) for recent changes. Auto-triggers on docs, documentation, changelog, readme, docstring, or jsdoc keywords. Asks the user which artifacts to produce and generates only the selected ones.
Use before any creative work — designing a feature, planning a change, building a component, adding functionality, modifying behavior, scoping a UI, or shaping new build work. Explores intent, constraints, and trade-offs, derives a topic slug, and ends by offering to hand off to spec-writing.
Reviews the current branch diff against the base branch, lists issues by severity (Blocker / Major / Minor / Nit), auto-fixes trivial issues (formatting, unused imports, dead code, comment typos), and flags everything else for the user. Auto-triggers on review / diff / PR / code-review / "review this" keywords. Works standalone on any branch in any repo; if a `docs/ultra-dev/<slug>/` directory maps to the current branch, additionally reads its `spec.md` and `plan.md` as extra context.
Generates an entity-relationship diagram (ERD) from a live database. Auto-triggers on database / schema / migration / ERD / diagram / table / relation / foreign-key keywords. Asks the user before introspecting; supports Postgres, SQLite, and SQL Server. Outputs erd.md (Mermaid for LLM context) and erd.html (rendered diagram for browser viewing) into the active feature directory.
Researches libraries, frameworks, services, APIs, and SDKs for an upcoming feature. Pulls live docs via the user-installed `context7` MCP server, writes a terse `research.md` to the active feature directory, and prints a short chat summary. Does NOT auto-trigger from generic prompts; runs only when the user invokes it directly or accepts the research offer at the start of `brainstorm`.
Produces `docs/ultra-dev/<slug>/spec.md` from a brainstorming conversation or existing context. Does NOT auto-trigger from generic prompts; runs only when chained from `brainstorm` or invoked explicitly via the Skill tool by name.
Post-hoc test generator. Auto-triggers on test, tests, testing, coverage, spec, unit test, integration test keywords. Reads files changed since the branch base, proposes a checklist of tests, and generates them on user confirmation. POST-HOC ONLY — no TDD enforcement, no failing-test-first workflow.