solopreneur

Android/Kotlin development expert. Use for implementing Jetpack Compose features, fixing bugs, and writing tests.

UI/UX design expert. Use for visual design, design systems, component styling, and design reviews — works across web, iOS, and Android.

iOS/macOS SwiftUI development expert. Use for implementing SwiftUI features, fixing bugs, writing tests, and architecture reviews.

AI Agent / LLM Application development expert. Use for implementing LangGraph workflows, multi-agent systems, tool calling, and streaming.

Next.js/React development expert. Use for implementing web features, fixing bugs, and writing tests.

Python development expert. Use for implementing FastAPI/LangGraph features, fixing bugs, and writing tests.

Web frontend development expert. Use for implementing React/Next.js features, fixing bugs, writing tests, and performance optimization.

Use when building Android apps with Jetpack Compose — covers @Preview setup (LocalInspectionMode, Vico charts), Scaffold + bottom nav + status bar insets, ModalBottomSheet nested-scroll jitter, ripple clipping on rounded corners, SwipeToDismissBox with transparent content, and locale-aware date formatting (MIUI quirks).

Multi-PR orchestration planner. Reads a task file, discusses PR splitting strategy with the user, produces plan + spec files, and schedules automatic execution. Use when the user wants to break a large feature into multiple PRs and have them implemented, reviewed, and merged automatically. Triggers: "autopilot", "orchestrate", "batch implement", "split into PRs and run", or when the user provides a todo/spec and wants unattended multi-PR execution.

Automated PR review loop that keeps running until the PR is clean. Triggers external reviewers (Codex bot, Codex CLI, Gemini), processes feedback, fixes issues, and re-triggers — repeating until no new suggestions remain. Supports `/greenlight external` to skip internal review and go straight to external reviewers, and `/greenlight external gemini` to specify a starting reviewer. Use when the user says "greenlight", "run reviews until clean", "get this PR approved", or wants automated review cycling on an open PR. Also auto-detects uncommitted mode: when on main branch with uncommitted changes and no PR, runs Codex CLI `--uncommitted` review loop only, fixes in-place without committing, until codex reports clean.

Generate a complete Go-To-Market strategy by analyzing a codebase and interviewing the user. Produces 4 strategy documents covering brand, market landscape, messaging, and channel playbook. Supports initial deep research (multi-session with interviews) and weekly incremental updates. Use when: "gtm", "go to market", "GTM plan", "brand strategy", "market analysis", "messaging framework", "generate GTM".

Remove AI writing patterns from text. Three modes: (1) pre-write — read before writing to internalize anti-AI-slop rules, (2) review — full post-writing audit with severity tiers and two-pass rewrite, (3) detect — flag-only audit without rewriting. Use when: editing drafts for AI tells, cleaning up AI writing, auditing copy, making text sound human, 去 AI 味, 潤稿, 檢查 AI 味. Also use as a pre-write reference before composing any public-facing text.

Use when building iOS/macOS apps with SwiftUI — covers localization (String Catalogs), date/time formatting, JSON date decoding, Previews, state management, sheet/navigation, list spacing, and animation patterns.

Session retrospective — reviews the current conversation to find mistakes, trace root causes, and propose process improvements. Use when the user says "session retro", "retrospective", "retro", or "review this session" at the end of a work session. Also useful mid-session after a correction to immediately analyze what went wrong.

Create brand-aware presentations using frontend-slides or reveal.js. Use when the user wants to make a presentation, slides, deck, or slideshow. Triggers: 做簡報、建 slides、presentation、投影片、做投影片、slide design、簡報設計。Always use this skill instead of frontend-slides or revealjs directly — it adds a brand setup step that saves time on color/asset iteration.

Tech-stack-aware expert code review using specialized subagents. Detects which tech stacks are in the diff, then dispatches the matching specialist agents (ios-dev, android-dev, llm-dev, python-dev, nextjs-dev, web-dev) to review against their skill-index best practices. Use when the user says "specialist review", "expert review", "stack review", or wants a multi-perspective code review with best practice verification. Also use after completing a significant implementation when thorough review is needed.

Periodic scanner for backlog and in-progress todos. Cross-references with GitHub PR status, reviews unhandled items, maintains worktrees, and notifies via Discord (if configured) or terminal output. Presents a confirmation checkpoint before executing actions. When the user approves an item, auto-implements in a worktree and runs the full review loop. Designed for /loop periodic execution. Use when the user says "babysit todos", "sweep backlog", "scan todos", "watch backlog", "what todos can I work on", or sets up a periodic loop (e.g., `/loop 24h /todos-babysit`).

Batch cleanup of backlog todos — checks which items are already completed or partially done by searching git history, then moves them to done/ or doing/. Use when the user says "cleanup todos", "triage backlog", "clean up backlog", "check todos status", or wants to classify backlog items based on actual implementation status.

Deep-review a single todo or spec file to evaluate problem clarity, solution quality, and implementation priority. Dispatches platform-specific expert agents for best-practice checks. Use when the user says "review this todo", "review todo", "evaluate this spec", provides a todo file path, or wants a go/no-go recommendation before implementing a backlog item.

Create a git worktree for a task and write a CONTEXT.md that captures the full problem context, so the next session can pick up without re-explanation. Use when the user says "open worktree", "worktree handoff", "create a worktree", "start a new branch for this", or wants to hand off a task to a fresh session with full context preserved.

LinkedIn organic growth consultant — diagnoses profiles, discusses goals, and co-creates a personalized 90-day growth plan. Backed by a research library covering the latest algorithm mechanics, content strategy, engagement tactics, and real case studies (updated 2026-04). Use when the user mentions LinkedIn growth, LinkedIn strategy, LinkedIn profile optimization, LinkedIn content plan, LinkedIn algorithm, building an audience on LinkedIn, getting more LinkedIn impressions/followers/leads, or wants to diagnose their LinkedIn presence. Also trigger when the user shares a LinkedIn profile URL or asks how to grow on any professional social platform.

Generate a product or company name through structured brief, multi-model candidate generation, and two-layer evaluation. Supports greenfield (new name) and rebrand (rename existing) modes. Grounded in Lexicon / Interbrand / Siegel+Gale methodology plus processing fluency, sound symbolism, and iconicity research. Auto-reuses `docs/gtm/` if present. Use when: "naming", "取名", "命名", "品牌名", "產品名字", "公司名字", "rebrand", "改名", "幫產品命名", "幫公司取名".

Switch perspectives to think through problems using the mental models of ten iconic thinkers: Elon Musk, Richard Feynman, Charlie Munger, Naval Ravikant, Steve Jobs, Nassim Taleb, Ilya Sutskever, Andrej Karpathy, Paul Graham, and Donald Trump. Each perspective is a distilled thinking OS based on deep research of primary sources, interviews, decision records, and external criticism. Use when the user says "perspective", "switch perspective", "think from X's viewpoint", "how would X see this", or mentions a specific thinker's core concept (e.g. "first principles", "antifragile", "leverage", "inversion", "focus means saying no", "cargo cult", "compression is understanding", "march of nines", "make something people want", "truthful hyperbole").

Trace when a bug was introduced, find the root cause commit, understand why it happened, and produce a structured post-mortem report. Use when the user says "post-mortem", "trace this bug", "when was this regression introduced", "find the root cause commit", "what commit broke this", or when they describe a crash/bug that appeared in a specific version. Works with any git repository.

Pre-implementation best practice review. Analyzes a technical plan against platform-specific best practices before coding begins. Use when the user says "preflight", "review the plan", "best practice review", "sanity check this approach", or wants to verify a technical plan is sound before implementation. Works best with context7 MCP server for documentation lookups (gracefully degrades without it).

Rebuild the per-machine extended skill index used by platform subagents. Scans installed Claude Code skills, classifies each by platform relevance (iOS, design, …) using the frontmatter description, and writes one file per platform under ~/.claude/solopreneur/skill-index/. Use when the user says "rebuild skill index", "refresh skill index", or after installing / removing platform-related skills.

Get an independent adversarial review of plans, specs, and design docs using Codex CLI as a second pair of eyes. Codex reads the document and related files fresh — no conversation context — then challenges it across 5 dimensions: Completeness, Consistency, Clarity, Scope, Feasibility. Use when the user says "second opinion", "challenge this plan", "adversarial review", "review this spec", or any request to get an independent outside perspective on a plan or spec. Works best with Codex CLI installed; falls back to a subagent-based review without it.

X/Twitter growth consultant — diagnoses profiles, discusses goals, and co-creates a personalized growth plan. Backed by a research library covering the latest algorithm mechanics, content strategy, engagement tactics, monetization, and real growth playbooks. Integrates with existing GTM strategy docs when available. Use this skill whenever the user mentions X/Twitter growth, profile optimization, tweet strategy, X audience building, follower growth, X content planning, or wants to diagnose/improve their X presence — even if they don't explicitly ask for a "growth plan." Also use when the user shares an X/Twitter profile URL or handle and wants feedback.

X/Twitter writing coach — helps you write single tweets, threads, and long-form posts. Generates hooks, suggests topics, reviews drafts, and explains craft principles grounded in Aesthetic Writing, RARE hooks, and the algorithmic reality of X. Use this skill when the user says things like "help me write a tweet," "review this tweet," "I need a thread idea," "make this hook better," "what should I post today," or shares a draft asking for feedback. For long-term account strategy (profile diagnosis, 12-week growth plans), use `x-growth` instead.