mdeapp

DEPRECATED — use the `linear` skill instead. "Manage Linear issues, projects, and teams"

DEPRECATED — use the `linear` skill instead. Create Linear issues. Use when creating bugs, tasks, or feature requests.

DEPRECATED — use the `linear` skill instead. Manage Linear documents. Use for creating and viewing documentation.

DEPRECATED — use the `linear` skill instead. Manage Linear labels. Use when creating, listing, or deleting labels.

DEPRECATED — use the `linear` skill instead. View Linear metrics. Use for velocity, burndown, and progress tracking.

DEPRECATED — use the `linear` skill instead. Manage project milestones - create, update, track target dates. Use when planning project deliverables.

DEPRECATED — use the `linear` skill instead. Linear project management - issues, projects, cycles, and roadmaps. Use for Linear-related tasks like managing issues, tracking sprints, and organizing projects.

DEPRECATED — use the `linear` skill instead. Create GitHub PRs linked to Linear issues. Use when creating pull requests, pushing code for review, or linking PRs to Linear issues.

DEPRECATED — use the `linear` skill instead. Manage Linear projects - full CRUD with labels, members, archive. Use when managing projects.

DEPRECATED — use the `linear` skill instead. Manage Linear issue relationships. Use for blocking, parent/child, duplicates.

DEPRECATED — use the `linear` skill instead. View Linear roadmaps. Use when viewing roadmap planning.

DEPRECATED — use the `linear` skill instead. Search Linear issues and projects. Use when finding issues, looking up bugs, or searching the backlog.

DEPRECATED — use the `linear` skill instead. Sprint planning and analytics - status, progress, burndown, velocity, carry-over. Use when managing sprints.

DEPRECATED — use the `linear` skill instead. View and manage workflow states. Use when checking or configuring issue statuses for a team.

DEPRECATED — use the `linear` skill instead. "Linear project management expert for issues, cycles, projects, and workflow automation"

DEPRECATED — use the `linear` skill instead. Manage Linear triage inbox. Use for unassigned issues needing attention.

DEPRECATED — use the `linear` skill instead. Update Linear issues. Use when changing status, priority, assignee, or labels.

DEPRECATED — use the `linear` skill instead. Manage custom views - create, list, apply saved views. Use when working with saved issue filters.

DEPRECATED — use the `linear` skill instead. Start/stop work on Linear issues. Use when beginning work, creating branches, or getting current issue context.

DEPRECATED — use the `linear` skill instead. mdeai Linear operations — create/update issues, cycles, milestones, and status updates for the sanjiovani workspace. Use whenever the user mentions Linear, SAN-###, Linear issues, cycles, milestones, initiative updates, project updates, phase:launch labels, or syncing tasks to Linear — even if they only say 'create the issue' or 'mark it Done'. Loads mdeai conventions from linear.md first, then delegates API/CLI mechanics to the generic linear skill. Owns executable issue descriptions: lettered completion-step checklists (A1…E5), Mermaid Gantt + sequence diagrams, persona-first plain English, and progress tracker sync (todo.md + changelog). Pair with mde-task-lifecycle (references/linear-issue-steps.md) and mermaid-diagrams (references/gantt.md). Do NOT use linear-automation or linear-claude-skill (archived).

Forensic verifier for mdeai task specs (tasks/core/F*.md) and audit reports (tasks/audit/*.md). Use whenever the user asks to "verify a task", "is this task 100% correct", "audit the audit", "is this safe to execute", "did this task really ship", "check if X is done", or before flipping any task from In Progress → Done — even if they only @-mention the task file. Also gate new task specs through the quality checklist before saving. Score specs with references/task-spec-rubric.md (spec score vs execution readiness). Prove every claim against disk + node_modules + MCPs; never trust a status field.

mdeai testing: Vitest, Playwright, browser MCP, smoke, CI, evals. Use routing table. NOT for: TDD-only workflow (test-driven-development), MCP server failures (troubleshooting), dedicated load-test design.

Use for worktrees, PRs, shipping, and cleanup. Enforces "small branch → focused PR → tested → merged → deployed → verified → next branch" — one worktree, one goal, one PR, merge gate (lint/build/test/floor/Vercel/prod), PR size limits, and forensic cleanup for messy state. Triggers include start worktree, open PR, split PR, audit/clean worktrees, one PR at a time, dirty worktrees.

Lean implementation loop for mdeapp (SAN-NNN) tasks. Use this skill whenever starting implementation of any Linear issue, choosing which tests to run, setting up or cleaning up a worktree, or when mde-task-lifecycle phases 3–4 feel heavy for the change at hand. Provides: the 7-step loop (read→implement→T1 test→typecheck→commit→push→PR), test-tier selection by change shape (T1 targeted ~2s / T2 domain ~5s / T3 floor ~3min), worktree bootstrap/audit/clean commands, local-vs-CI verification matrix, and structured evidence capture. ALWAYS invoke this before running `npm run floor` locally, spinning up the dev server for a backend-only change, or doing a full-suite test run after a single-file edit. Replaces the "lint-per-chunk + build-per-chunk + floor-before-commit" pattern that burns 2–3min per task cycle with no safety benefit.

Mastra framework: docs lookup (links.md, mastraDocs MCP, embedded docs), agents, workflows, streaming, browser, tools, memory, RAG, processors, CopilotKit guide. Verify from installed docs — never trust training-data APIs. NOT for: product intent routing (mastra-routing), Managed Agents API harness (mde-agents), non-Mastra frameworks. Load when editing src/mastra/**, Mastra tools/workflows, memory, streaming/events, browser automation, or Mastra doc URLs.

mdeai Google Maps: Places (New), ChatMap, Gemini grounding, MCP tools, keys, Mastra maps. Quick routing; legacy `google-maps` name archived (see `_archive/2026-05-14`). NOT for: Mapbox-only, pure GIS math, non-Google stacks, generic off-repo tutorials.

Orchestrator for real-estate work on mdeai.co — V1 marketplace product, V2 MLS technical playbook, industry context, plus 3 specialized sub-agents (lead qualifier, neighborhood guide, property description). Use when building real-estate features, writing listings, qualifying leads, generating neighborhood guides, or making product/architectural decisions for the real-estate vertical. Triggers: real estate, MLS, listing, property, broker, agent, neighborhood guide, lead qualification, property description, mdeai real estate. Do NOT use for: short-term rental flows that live in `mde-task-lifecycle` or generic CRM tasks.

mdeai Supabase: DB, RLS, Auth, Edge Functions, Realtime, Storage, vectors, migrations, pgvector. Use for ANY Supabase work on this repo — schema changes, RLS policies, edge function deploy/debug, auth.uid(), service role, supabase-js, Realtime channels, Storage buckets, or Supabase MCP. Load topic files + references/project-rules/. NOT for: Auth0/Clerk-only auth, raw Postgres without Supabase, non-Supabase BaaS.

Five-phase orchestrator for mdeai.co tasks — plan, research, implement, test, ship. Routes each phase to specialist skills (mde-writing-plans, testing for Vitest/Playwright, `/deploy-check` for pre-ship gates, mdeai-executor) and owns boundaries between them. Use when the user wants to work on a task by ID (work on task 17A, implement 18B, ship this prompt, process the backlog, next P0, go through the lifecycle, close out a task, run a task), or asks to move anything in tasks/prompts/ from spec to shipped. Does NOT write PRDs (use mde-writing-plans), debug arbitrary bugs (use systematic-debugging), or author new skills (use mde-prompting). Replaces overlapping ad-hoc skills create-tasks, spec-tasks, tasks, tasks-generator, prd-taskmaster, task-prd-creator, executing-tasks.

Orchestrator for Vercel platform work — deploy operations (`vercel deploy`, preview/production, rolling releases, environment variables, domain config, deployment troubleshooting) and React/Next.js performance best practices from Vercel Engineering (Server Components, data fetching patterns, bundle optimization, caching, ISR, streaming, Core Web Vitals). Use when deploying to Vercel, troubleshooting a deployment, configuring envs/domains, or writing/reviewing/refactoring React/Next.js code for performance. Triggers: vercel, vercel deploy, preview deployment, production deployment, push live, vercel.ts, vercel.json, vercel env, react performance, next.js, server component, RSC, ISR, streaming, bundle size, core web vitals. Do NOT use for: non-Vercel hosting (Netlify, Cloudflare Pages) or React patterns unrelated to performance.

TEMPLATE — not a real skill. Copy this folder to `.claude/skills/mde-<domain>/` and remove `disable-model-invocation`. One-line WHAT + WHEN + at least 6 trigger phrases users would actually say. Trigger when the user says "phrase A", "phrase B", "phrase C", "phrase D", "phrase E", or "phrase F". Does NOT handle <out-of-scope topic 1> (use <other-skill>) or <out-of-scope topic 2> (use <other-skill>). Replaces the legacy skills <list>.

Create or modify forms in admin or vendor dashboards in the Mercur basic starter using stable form structure, validation, submission guards, and dashboard UI conventions.

Create or modify pages in admin or vendor dashboards in the Mercur basic starter using correct routing, page composition, i18n, and dashboard UI conventions.

Create or modify tabbed workflows in admin or vendor dashboards in the Mercur basic starter using stable tab structure, validation scope, and i18n-aware wizard conventions.

Keep custom admin and vendor UI in the Mercur basic starter aligned with local wrappers and @medusajs/ui. Use when adding reusable UI, interactive primitives, overlays, menus, selectors, or custom dashboard components.

Discover, evaluate, install, and verify Mercur blocks in the basic starter using `blocks.json` aliases and CLI workflows.

Use Mercur CLI commands correctly inside a project created from the Mercur basic starter. Use when choosing between `create`, `init`, `add`, `search`, `view`, and `diff`.

Analyze a Mercur 1.x project and guide migration to 2.0. Self-contained — works without access to the mercur monorepo.