outworkos_diy

Bootstraps a new project directory with base configuration files, MCP servers, API connections, and template skills from OutworkOS. Use when setting up a new project, initializing a workspace, or configuring MCP integrations for a new directory.

Gain-based weekly reflection. Looks backward at what moved forward across all projects, written as narrative prose. Modeled on The Gap and The Gain — measures progress against where you were, not where you wish you were.

Composes, drafts, and sends emails via Gmail. Enforces natural writing style (no em dashes, no AI tells), proper threading, and correct signature handling. Use whenever drafting, replying to, or sending any email.

"Produces a meeting prep briefing for upcoming calendar meetings by researching attendees, email history, past meetings, and company backgrounds. Use when the user asks to prepare for meetings, prep for tomorrow, review upcoming meetings, get ready for a meeting day, or asks 'what meetings do I have'. Also use when the user asks to research a specific meeting or contact before a call."

"Cross-project inbox scan — cleans noise, routes signals to projects, creates Todoist tasks for gaps, and ranks priorities by time block. Replaces /triage, /inbox-zero, and /sitrep."

Generates Spotify playlists using MCP spotify tools. Use when user wants to create a playlist based on mood, genre, similar artists, or any theme. Handles auth, track discovery, playlist creation, and adding tracks.

Git Branch Cleanup

Creates a project context map by interviewing the user, exploring connected data sources, and scaffolding project configuration (CLAUDE.md, settings.json, custom agents). Use when starting a new project or needing to map all systems of record for a project.

Guides users through Jobs-to-be-Done discovery to define feature requirements before implementation. Use when user wants to define, document, or plan a new feature or enhancement. Interviews for problem and jobs, analyzes codebase, produces TDD-focused implementation plan ready for GitHub issue creation.

Document as GitHub Issue

Creates comprehensive GitHub issues that serve as source of truth for changes. Use when user asks to create a GitHub issue for a bug fix, feature, update, or any other change. Captures all research, analysis, and implementation context for the next developer.

Creates persistent, recurring local automations on macOS using launchd. Interviews for requirements, creates wrapper scripts and plists, loads the job, tests it, and updates the project context map.

"Records project events and working sessions to log.md. Default mode captures end-of-session accomplishments. 'refresh' mode queries data sources for external events. Use at end of session, when user says 'log this', or to pull in external timeline events."

Sends a Pushover push notification to Matt when Claude is blocked and needs user input to continue working. Use PROACTIVELY when stuck on a decision, missing information, or needing approval during an active task. Do NOT use for routine questions at the end of a task.

Enforces DevOps best practices for commits, merges, and PRs. Use when user requests a commit, merge, or pull request. Analyzes changes, validates quality, composes conventional commit messages, and ensures user approval before any git operations.

Analyzes web applications for mobile responsive design issues including touch targets, typography, breakpoints, navigation, images, and performance. Provides file-by-file tactical recommendations and implements fixes for Next.js 15, Tailwind CSS 4, Radix UI, and shadcn/ui projects. Use when auditing mobile UX, optimizing responsive layouts, or ensuring mobile-first best practices.

Initializes project memory by scanning the current project structure, asking targeted questions, and writing structured memory records to the database. Use via /memory-init when starting work on a new project.

Core memory system definition. Defines the persistent memory schema, operations, and self-modification rules. All memories stored in the Supabase `memories` table. Referenced by all other memory skills and agents.

Retrieves and presents relevant memories from the database for the current project context. Queries the memories table and returns a synthesized summary. Use via /recall at the start of a session or when context is needed.

Extracts and persists memories from the current conversation to the database. Scans for decisions, preferences, patterns, issues, and learnings, then dispatches the memory-extractor agent. Use manually via /rem-sleep or triggered automatically before context compaction.

Guides product requirements definition using SVPG principles. Use when defining product requirements, transforming feature requests into problems, or assessing product opportunities. Focuses on problem-first thinking, four-risks assessment, and separating problems from solutions.

Pragmatic QA system that analyzes codebases, designs risk-proportional testing strategies, and manages test suites built with TDD. Activates when user requests QA analysis, test strategy, test suite audit, or testing guidance. Makes evidence-based recommendations for testing rigor matched to actual business and technical risk.

Use after running tests to clean up test data. Analyzes tenant structure, identifies test-created tenants (empty shells), and safely removes only records with "Test" pattern. Always shows preview and requires confirmation before deletion.

Senior developer specialist with expertise in Node.js, React, TypeScript, and Supabase. Provides on-demand architectural analysis, refactoring guidance, and performance optimization. Focuses on Supabase (40%), React/Performance (30%), TypeScript (20%), and Node.js/Next.js APIs (10%). Uses context7 for documentation-backed recommendations.

Interview-based blog post creation that shares hard-won lessons through storytelling. Use when crafting posts about personal learnings, best practices, or experience-based insights.

Drafts and publishes LinkedIn posts via the official API. Supports text, images, and link previews. Handles first-time OAuth setup and token refresh. Use when drafting, reviewing, or publishing LinkedIn posts.

Processes a Fireflies meeting recording into a structured summary with Todoist action items and a database log entry. Use when processing meeting notes, creating tasks from meetings, or logging a recorded call.

"Reviews all conversations captured by the Limitless Pendant for a given day, cross-references with Calendar events and Fireflies transcripts to pick the best source, then collaboratively matches signals to existing projects, contacts, and ideas. Use when processing daily pendant data, reviewing ambient conversations, or capturing insights from in-person meetings."

Run and verify the daily receipt-to-Drive pipeline locally. Scans Gmail for purchase emails, uploads to Google Drive, and emails the configured accounting address.

/requirements - Define Feature Requirements

Interactive walkthrough of the top 5 risk-labelled Todoist tasks. Pulls full context (emails, meetings, task history) for each item and helps resolve them one by one — reply, close, reschedule, or delegate. Use after /risk to work through flagged items.

"Cross-project risk analyzer — identifies open Todoist tasks where follow-through is missing by cross-referencing sent mail, project logs, and calendar. Scores risk by temporal compounding, escalates Todoist priorities for Critical/At Risk items, and sends Pushover alerts. Use daily, when asking 'what have I dropped', or to audit follow-through across projects."

Fetches a LinkedIn profile via ScrapingDog API and displays a formatted summary (name, role, location, followers, about, experience, education). Use when asked to look up, scrape, or fetch a LinkedIn profile for a speaker, contact, or person. Accepts a LinkedIn profile URL or ID.

/ship-it - Commit, Push, and Create PR

/start-work - Begin Working on a GitHub Issue

Drafts articles on Substack from markdown or text files. Converts content to Tiptap JSON, authenticates via Chrome browser tools, and creates drafts via Substack API. Use when publishing, drafting, or posting content to Substack.

Use when implementing any feature or bugfix: write the test FIRST, watch it fail, then write minimal code to pass. For complex work, creates an epic with phased issues. Ensures tests verify behavior by requiring failure before implementation.

Reference for Todoist Sync API task management — reading, creating, updating, moving, and completing tasks using the sections-based status model (Triage/Do/Waiting/Defer). Use when managing Todoist tasks, creating batch task operations, checking task status, or building Todoist API payloads. Not a user-invocable skill.

Audits Supabase Vault secrets — lists all stored secrets, tests API token validity, checks OAuth token freshness, and reports missing or expired credentials. Use when troubleshooting auth failures, after setup, or for periodic health checks.

Syncs Todoist, Gmail, Calendar, and Fireflies to surface untracked action items and recommend the next task to work on. Use when starting a session, asking 'what's next', or checking priorities for any project.