Menü
MCP server usage for Dart MCP, Firecrawl, Hugging Face. Tool priorities, commands, workflows.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
Development process rules. Use when planning complex tasks, implementing incrementally, managing packages, or following communication guidelines.
Dart 3 patterns including switch/if-case, pattern matching, sealed classes, and records. Use when writing exhaustive switches, destructuring, pattern matching on types, or returning multiple values with records.
Core code quality rules for Dart/Flutter. Use when formatting, handling async/null safety, organizing classes, or following Dart best practices.
Code review checklist for PRs/MRs. Use when reviewing pull requests, examining code changes, or when the user asks for a code review.
Data handling patterns for Dart/Flutter. Use when implementing serialization, caching, layer separation, DTOs, repositories, offline-first, or data validation.
Dependency injection patterns for Dart/Flutter applications. Use when configuring get_it, registering dependencies, using constructor injection, setting up composition root, and testing with mocks.
| name | cursor-mcp-servers |
| description | MCP server usage for Dart MCP, Firecrawl, Hugging Face. Tool priorities, commands, workflows. |
| Task | Tool | Fallback |
|---|---|---|
| Dart/Flutter packages | pub_dev_search | Firecrawl scrape pub.dev |
| Code analysis/format/tests | Dart MCP | — |
| Web content | Firecrawl scrape/map | web_search |
| ML models/datasets/papers | Hugging Face MCP | Firecrawl |
firecrawl_search: UNAVAILABLE (503) — use web_search or map + scrape.
Setup: add_roots required before workspace commands. Format: [{uri: "file:///path", name: "optional"}].
Commands:
Errors: Verify SDK; DTD needs fresh URI after reconnect; dart_fix before manual fixes; check network for pub.
scrape — single page (formats: markdown/html; maxAge for caching)batch_scrape → check_batch_status — multiple URLsmap — discover URLs on site, then scrapecrawl → check_crawl_status — full site (caution: token overflow; limit depth/pages)extract — structured data with schema/promptConfig: FIRECRAWL_API_KEY required. maxAge for caching (172800000 = 2 days).
Errors: onlyMainContent: true for cleaner scrape; JS-heavy sites may fail.
model_search (query, sort, limit, author, library, task) — max 100/querydataset_search (query, sort, tags)paper_search (query, results_limit, concise_only) — concise_only for broad searchspace_search, hub_repo_details (repo_ids author/name, repo_type) — max 10/callhf_doc_search → hf_doc_fetch (offset for large docs)gr1_flux1_schnell_infer (prompt, width, height, num_inference_steps) — defaults: 4 steps, 1024x1024Errors: hf_whoami for auth; specify repo_type if auto-detect fails; filter early.
Package info: pub_dev_search(query: "package_name")
Add dependency: pub(command: "add", packageName: "x", roots: [{root: "file:///path"}])
Analyze project: add_roots → analyze_files → dart_fix → dart_format
Web docs: web_search for URLs → scrape(url, formats: ["markdown"], onlyMainContent: true)
Batch scrape: batch_scrape(urls) → check_batch_status(id)
ML models: model_search with filters → hub_repo_details
DTD: connect_dart_tooling_daemon(uri) → get_widget_tree / hot_reload