Plan before you build. Ritual runs a structured exploration — discovery questions, recommendations, a validated build brief — so your coding agent gets the intent, constraints, and prior decisions it can't infer from the repo. Bundles the Ritual MCP server connection and the /ritual skill.
Log food, calories, macros, and weight to a personal Notion database by talking to Claude. Estimates portions and macros from plain-language descriptions, records how confident each estimate is, and reports daily totals and multi-day trends. Setup is a single Notion connector sign-in — no API keys or credential files. Readable and writable from Claude Code, claude.ai, and the Notion mobile app.
Bundled review and diagram playgrounds: /plan-review, /design-review, /architecture-map, /code-diagram, and a standalone /devserver launcher. Single install with one shared devserver binary and PTY bridge for session-resumed feedback delivery.
Manage the full GitHub issue lifecycle — from picking up an issue through branch setup, implementation, and raising a PR. Includes /swarm (agent-driven) and /swarm-exec (harness-driven) agentic flows.
Productivity tools for Claude Code workflows. The /goal skill sharpens a vague intent into a precise, self-checkable completion condition through a focused interview (one question at a time, max three), then hands it off to Claude Code's built-in /goal command. Includes /ui-first — a dual-mode UI-structure designer/auditor that derives navigation from the user's work, never the domain model.
Your storekit account in Claude: sales, takings, orders, refunds, menus, guests and reviews, plus approved snoozes, discount codes and payment links. Sign in with your storekit dashboard login.
A harness for a personal agent: identity, law, memory and a self-extending skill library in one git repo, wired into every Claude Code and Codex session and updated without ever overwriting what you wrote.
Technical program management: /tpm charters a program, builds a buffered plan, logs risks with ordered strategies, sets up stakeholder communication and RACI, and writes audience-leveled status reports. Frameworks are bundled locally; no RAG or network access at runtime.
Prefer existing code, the standard library and native platform features; keep the smallest correct implementation without dropping validation, security or accessibility.
Audits a frontend project against the 30 Laws of UX (lawsofux.com). Detects the UI stack, scores each applicable law from concrete source-code signals, and produces a graded report (0-100 + A-F) with a findings table, a coverage band, and a prioritized quick-wins-vs-larger-efforts fix list. Static by default; an optional --url pass measures the running app. Read-only unless you run /ux-fix.
Financial-workflow skill series: configurable market/account data sources (MCP, direct HTTP, or local Python), portfolio/account reporting and DCF valuation bridged into anthropics/financial-services, standalone strategy backtesting, a market-wide TWSE institutional-buying screen, and a shared Finnhub REST query skill.
Make sessions continuous. Wrap one up into a handoff the next one loads, start with today's date and a compact build-loop brief, warn when another session is already working here, and put cost, context and connected tool health in the status line.
Bridge to Slack's web API via browser session tokens (xoxc/xoxd) — surfaces what the hosted Slack MCP can't: true unread (Activity feed) with real bulk mark-read, and the saved-for-later 'Later' list (list/complete/snooze/archive). Skills: /slack-triage (interactive unread board), /slack-saved (grounded, stateful followup worklist with a durable decision store), /slack-unwrapped (stats recap), plus /slack-setup and /slack-doctor. Bundles a custom Python MCP server (13 tools) + a Chrome token-grabber extension. Run /slack-setup once.