Claude Skills by harshadmadaye

Turns a rough idea into an approved design or written spec through one-question-at-a-time dialogue, scaled to spike, bounded, or architectural work. Use when asked to create a feature, build a component, add functionality, modify behavior, or start a new project, before any code is written.

Runs a multi-agent debate over a project idea. Four angle agents (product, engineering, skeptic, market research) propose and rebut, one judge ranks features keep / improve / add / cut with a phased plan, one designer turns the verdict into flows, screens, data and architecture, and the result becomes a markdown spec plus a design artifact like jump-start's. Use when the developer runs /opm:brew-idea <brief>, says "brew this idea", asks to brainstorm a project with several agents, or wants a ...

Captures a solved, non-obvious problem as a durable learning doc under docs/solutions/ with validated frontmatter, and retrieves matching learnings before new work starts. Use after a tricky bug fix, a surprising root cause, or a decision with real tradeoffs, and at the start of planning to check whether the team already learned this.

Executes a written implementation plan task-by-task by dispatching a fresh implementer subagent per task, a reviewer subagent after each, and a progress ledger that survives context compaction. Use when a plan exists in docs/plans/ and the user wants it implemented, resumed after interruption, or finished with verification and a PR.

Turns a spec or document into a narrated explainer video. Produces a 1920x1080 MP4 of illustrated still slides with a neural voice-over and a sidecar .srt; it is a narrated slideshow with fades, not animation. Slides come from a script-rendered layout kit so the run is cheap and repeatable on macOS, Linux and Windows. Use when the developer runs /opm:explainer-video <path>, asks for an explainer or walkthrough video of a spec, brew-idea result, milestone summary or other document, or wants to...

Provides production Flutter and Dart 3 patterns covering widget composition, state management choice (Riverpod, BLoC, plain ChangeNotifier), immutability with freezed and records, sealed-class state, async and streams, Firebase usage (Auth, Firestore, Functions), navigation, and widget/integration testing. Use when writing, reviewing, or architecting Flutter apps or Dart packages.

Creates an entire new project from one prompt. Takes a project name plus a brief, clarifies only what would change the architecture, publishes a design artifact and re-confirms after every revision, then runs an unattended multi-agent OPM build in Claude Code's auto-accept permission mode, verifies it, and leaves the app running locally or on a device or simulator. Use when the user runs /opm:jump-start <project-name> <brief> or asks to build a whole new project from a description.

Structures multi-week work as a milestone with a roadmap, a short always-read STATE digest, phased executable PLAN files with YAML frontmatter and XML task bodies, wave-based parallel execution by fresh subagents, deviation rules, checkpoints, and SUMMARY files for cheap history. Use when scope exceeds about one week or eight tasks, or when starting a greenfield project.

Provides idiomatic modern Python (3.11+) patterns covering type hints, dataclasses and pydantic at boundaries, pathlib, context managers, generators, async basics, explicit error handling, and a uv/ruff/pytest project layout. Use when writing, reviewing, or refactoring Python code or setting up a Python project.

Provides idiomatic React 18/19 and Next.js App Router patterns covering hooks discipline, composition, state colocation, server/client boundaries, Suspense and error boundaries, forms, data fetching, list keys, effects hygiene, and measurement-first performance work. Use when writing, reviewing, or refactoring React components, custom hooks, or Next.js routes.

Old name for opm:explainer-video, kept for one release so existing habits keep working. Use when the developer runs /opm:story-video <path>.

Drives test-driven development through a strict RED-GREEN-REFACTOR cycle with test-runner detection, a "fails for the right reason" gate, checkpoint commits, and a short evidence report. Use when implementing a feature, fixing a bug, or refactoring logic where the expected behaviour can be written as a test before the code exists.

Establishes how to find and invoke OPM skills before taking any other action, including clarifying questions. Use when starting any conversation or task, before exploring the codebase, answering a question, entering plan mode, or writing code.

Requires running verification commands and reading their actual output before any claim that work is complete, fixed, or passing. Use when about to say done, tests pass, bug fixed, or build works; before committing, creating a PR, marking a task complete, or trusting a subagent's success report.

Runs the full quality gate loop (build, types, lint, tests, secrets scan, diff review) for Node, Python, and Flutter projects and produces a pass/fail report. Use after finishing a feature or refactor, before opening a PR, or whenever a session is about to claim that code is complete.

Writes a step-by-step implementation plan of bite-sized TDD tasks with exact file paths, interfaces, test code, and commits, from an approved spec. Use when a design or spec has been approved for a multi-step task and before touching any code.