.cursor/skills/toolbelt/SKILL.md
Canonical reference for all MCP servers, CLI tools, and platform integrations available in this workspace. Use when any skill or workflow needs to call Datadog, Supabase, Vercel, Figma, GitHub, Context7, or the browser MCP, or when starting a local dev server. Read this before using any integration tool.
npx skillsauth add Columbia-Cloudworks-LLC/EquipQR toolbeltInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Single source of truth for every integration available to the agent in this workspace. Other skills should reference this skill instead of documenting tool usage themselves.
| Integration | MCP Server | Primary Use |
|---|---|---|
| Datadog | plugin-datadog-datadog | Observability: logs, metrics, APM, RUM, monitors, incidents |
| Supabase | plugin-supabase-supabase | Backend: migrations, edge functions, SQL, tables, branches |
| Vercel | plugin-vercel-vercel | Frontend hosting: deployments, build logs, runtime logs |
| Figma | plugin-figma-figma | Design: read designs, write to canvas, Code Connect |
| Context7 | plugin-context7-plugin-context7 | Documentation: up-to-date library/framework docs |
| Browser | cursor-ide-browser | Testing: navigate, snapshot, interact with live pages |
| GitHub | gh CLI | Work tracking: issues, PRs, org project board |
Server: plugin-datadog-datadog
Every tool requires a telemetry object with an intent string describing why the tool is being called. Never include PII or secrets in the intent.
| Tool | When to Use |
|---|---|
| search_datadog_logs | Raw log search (requires query) |
| analyze_datadog_logs | SQL aggregation on logs (requires sql_query) |
| search_datadog_services | Discover services, teams, links |
| search_datadog_spans | Raw APM trace spans |
| aggregate_spans | Aggregate APM data (counts, latency percentiles) |
| search_datadog_monitors | Monitor status and alerts |
| search_datadog_dashboards | Find dashboards by name |
| get_datadog_metric | Timeseries metric queries |
| search_datadog_rum_events | Real User Monitoring raw events |
| search_datadog_incidents | Incident search by severity/state |
| get_widget | Render dashboard widget data |
search_* for raw events; use aggregate_* or analyze_* for counts and aggregations.check_datadog_mcp_setup diagnoses permission issues.Server: plugin-supabase-supabase
| Key | Value |
|---|---|
| Production project ID | ymxkzronkhwxzcdcbnwq |
| Organization ID | wjoecvchliqslvoukmpg |
| Tool | When to Use |
|---|---|
| list_migrations | Audit applied migrations |
| apply_migration | Apply DDL migration (requires name + query) |
| execute_sql | Run read queries or DML (not DDL) |
| list_tables | Inspect schema (schemas, verbose params) |
| list_edge_functions | Inventory deployed edge functions |
| get_edge_function | Read edge function source |
| deploy_edge_function | Push new edge function version |
| get_logs | Tail logs by service (api, postgres, edge-function, auth, storage, realtime) |
| get_advisors | Security and performance advisories |
| generate_typescript_types | Regenerate DB types |
| list_branches | Dev branch status |
| search_docs | Search Supabase documentation |
project_id: "ymxkzronkhwxzcdcbnwq" for production queries.apply_migration for DDL; execute_sql for SELECT/DML.files array with name + content per file.Server: plugin-vercel-vercel
| Key | Value |
|---|---|
| Team ID | team_78VeGDURoofThjZNJOKEBpP5 |
| EquipQR project ID | prj_P9hRun4B2OdGy8ACCnb0f7jNG6UA |
| Tool | When to Use |
|---|---|
| list_deployments | Recent deployments with status and commit info |
| get_deployment | Single deployment detail by ID or URL |
| get_deployment_build_logs | Build output for a specific deployment |
| get_runtime_logs | Runtime logs (filter by env, level, status) |
| get_project | Project configuration |
| deploy_to_vercel | Trigger a deployment |
| list_toolbar_threads | Vercel toolbar comment threads |
| web_fetch_vercel_url | Fetch protected deployment URLs |
| search_vercel_documentation | Search Vercel docs |
teamId: "team_78VeGDURoofThjZNJOKEBpP5".projectId — use the constant above for EquipQR.Server: plugin-figma-figma
| Key | Value |
|---|---|
| Plan key | team::1523012625938464812 |
| Tool | When to Use |
|---|---|
| get_design_context | Primary design-to-code tool (returns code + screenshot) |
| get_screenshot | Visual snapshot of a Figma node |
| get_metadata | XML structure overview of a node |
| use_figma | Run Plugin API JavaScript for create/edit/inspect |
| search_design_system | Search design libraries for components, variables, styles |
| get_variable_defs | Design token definitions |
| generate_figma_design | Import HTML/web page into Figma |
| generate_diagram | Create FigJam diagrams from Mermaid |
| whoami | Debug auth and permissions |
Extract fileKey and nodeId from Figma URLs:
figma.com/design/:fileKey/:fileName?node-id=:nodeId — convert - to : in nodeIdfigma.com/design/:fileKey/branch/:branchKey/:fileName — use branchKey as fileKeyfigma-use skill before calling use_figma.get_design_context is the default starting tool for design-to-code work.get_figjam instead of get_design_context.Server: plugin-context7-plugin-context7
| Tool | When to Use |
|---|---|
| resolve-library-id | Convert a library name to a Context7 ID (required first step) |
| query-docs | Fetch docs/examples for a resolved library ID |
resolve-library-id before query-docs unless the user provides a /org/project ID directly.query-docs more than 3 times per question.Server: cursor-ide-browser
browser_navigate to a URL (reuses existing tab by default)browser_snapshot to get the page accessibility tree and element refsbrowser_click, browser_fill, browser_type, browser_select_option, etc.browser_take_screenshot for visual verificationbrowser_lock / unlock to prevent user interference during automation| Tool | When to Use |
|---|---|
| browser_navigate | Go to a URL |
| browser_snapshot | Get page structure and refs for interaction |
| browser_take_screenshot | Visual screenshot |
| browser_click | Click elements by ref |
| browser_fill | Replace input field content |
| browser_type | Append text to input |
| browser_console_messages | Read browser console |
| browser_network_requests | Inspect network activity |
| browser_search | Ctrl+F text search on page |
| browser_scroll | Scroll page or element |
| browser_profile_start / _stop | CPU profiling |
browser_snapshot before interacting — refs are tied to the latest snapshot.browser_mouse_click_xy), take a fresh screenshot immediately before.browser_lock only after a tab exists. Correct order: navigate → lock → interact → unlock.CLI: gh (GitHub CLI, authenticated as viralarchitect)
| Key | Value |
|---|---|
| Repository | Columbia-Cloudworks-LLC/EquipQR |
| Org project (EquipQR board) | #5 (ID: PVT_kwDOCKvMq84BOBTY) |
| Org project (Legislative Red Team) | #6 |
# Issues
gh issue list --repo Columbia-Cloudworks-LLC/EquipQR --limit 10 --json number,title,state
gh issue view 123 --repo Columbia-Cloudworks-LLC/EquipQR
# Pull requests
gh pr list --repo Columbia-Cloudworks-LLC/EquipQR --json number,title,state
gh pr view 123 --repo Columbia-Cloudworks-LLC/EquipQR
# Org project board
gh project list --owner Columbia-Cloudworks-LLC --format json
gh project item-list 5 --owner Columbia-Cloudworks-LLC --format json
# PR review comments
gh api repos/Columbia-Cloudworks-LLC/EquipQR/pulls/123/comments
--json for structured output; avoid --jq with complex expressions in PowerShell (escaping issues).read:project scope; mutating items needs project scope.-i (interactive) flags.Never start the dev server directly. The user must run:
.\dev-start.bat # standard start
.\dev-start.bat -Force # force restart
This script injects secrets from 1Password into .env files. The agent cannot access 1Password, so starting the server without the user present will hang waiting for authentication.
If the app is partially running, the user should run .\dev-stop.bat first.
dev-start.bat.npm run dev or npx vite directly.| I need to... | Use |
|---|---|
| Check production logs or errors | Datadog search_datadog_logs |
| See what's deployed | Vercel list_deployments |
| Check build failures | Vercel get_deployment_build_logs |
| Run a database query | Supabase execute_sql |
| Check migration state | Supabase list_migrations |
| Inspect edge function code | Supabase get_edge_function |
| Look up library API docs | Context7 resolve-library-id → query-docs |
| Test the live app visually | Browser browser_navigate → browser_snapshot |
| Read a Figma design | Figma get_design_context |
| Write to Figma canvas | Figma use_figma (read figma-use skill first) |
| Check open issues/bugs | GitHub gh issue list |
| See project board status | GitHub gh project item-list 5 |
| Monitor frontend performance | Datadog search_datadog_rum_events or aggregate_rum_events |
| Check runtime errors | Vercel get_runtime_logs |
| Get security/perf advisories | Supabase get_advisors |
development
React performance optimization guidelines from Vercel Engineering, with EquipQR-specific mappings (Vite + React Router + TanStack Query). Use when writing, reviewing, or refactoring React code in this repo, especially around waterfalls, bundle size, and re-renders.
testing
Postgres performance optimization and best practices from Supabase, adapted to EquipQR's Supabase (Postgres + RLS) workflow. Use when editing SQL, migrations, indexes, or RLS policies.
development
Applies EquipQR's brand colors and design-system tokens to any artifact that should match EquipQR's look-and-feel. Use it when brand colors, style guidelines, visual formatting, or EquipQR design standards apply.
development
Use when auditing dependency health, API contract consistency, shared data shapes, or brittle integration seams between modules, services, and packages.