Coding standards, frameworks, refactoring, and code review (16.7K skills)
Upgrade the Anthropic Python or TypeScript SDK and adapt applications to API changes such as moving from Text Completions to Messages. Review breaking changes, update request shapes and streaming code, run compatibility checks, and keep a rollback path for the release.
Add safeguards to Claude API integrations, including circuit breakers that pause failing calls, bounded retries, timeouts, idempotency, and approved fallbacks. It helps test recovery behavior while keeping prompts, responses, and secrets out of telemetry.
Choose and plan a Claude API architecture for interactive requests, background jobs, or model routing. The skill lays out service components, failure handling, and safeguards so a team can test a selected pattern before production use.
Manage Claude API request and token budgets with coordinated rate controls. Handle 429 errors using response headers and bounded retries, and organize queued or batch work to fit available capacity across application instances.
Plan a move from OpenAI, Gemini, or another LLM provider to the Claude API. Map request and response fields, prompts, and tool definitions, and structure a provider-neutral adapter. Compare providers with synthetic fixtures and prepare a canary or rollback plan.
Set up a local development loop for Claude API projects with a reusable client, mock responses, request logging, and hot reload. Run routine tests without API calls, track model and token usage, and enable live requests deliberately when checking real API behavior.
Create small working examples for the Claude Messages API in Python or TypeScript. Use them to check credentials and learn how to send text, analyze an image, stream output, and inspect response metadata.
Use these TypeScript patterns to build Apify Actors with Crawlee or call Actors through apify-client. Examples cover crawler choice, datasets and key-value stores, proxies, typed wrappers, and handling failed requests without stopping a crawl.
Set up Apollo.io API access in TypeScript or Python by configuring an HTTP client, API key headers, and environment variables. Verify authentication with a health check and troubleshoot missing keys, insufficient permissions, rate limits, and network errors.
Design Apify data collection systems using a standalone scraper, a multi-stage pipeline, or an application integration. Get directory layouts, TypeScript skeletons, input validation, configuration patterns, and health checks suited to the chosen structure.
Plan and carry out a controlled migration of an Algolia JavaScript client from v4 to v5. Inventory package versions and API usage, update method calls and waits against the installed types and official migration guidance, run compatibility checks, and document remaining gaps and rollback steps.
Choose an architecture for Claude API workloads, from serverless functions and streaming microservices to queued batch jobs and multi-model routing. Receive trade-offs, scaling and failure controls, and a validation and rollout plan suited to the workload.
Investigate slow queries in PostgreSQL, MySQL, and MongoDB using execution plans, which show how queries run, and performance metrics. Produce index suggestions, query rewrites, and configuration recommendations ranked by likely impact and implementation effort.
Improve Anima design-to-code generation with version-aware caching and regeneration of changed components. Measure processing time, output size, and validation results to assess performance gains while preserving design fidelity, accessibility, and build requirements.
Assess whether an Anima and Figma design-to-code pipeline is ready for production. Review credentials, source traceability, generated-code and visual checks, monitoring, and rollback evidence, then produce a go/no-go record for release owners.
Manage rate limits when generating code from Figma designs with Anima. Set bounded waits and retries, track successful and failed nodes, and resume only the failed work, using the account’s documented quota instead of assumed limits.
Build Claude API workflows for live, incremental responses or asynchronous bulk processing. Handle streaming events and collect final usage details, or submit up to 100,000 requests through the Message Batches API and retrieve results by custom ID.
Build Claude assistants that call functions or APIs through the Messages API. Define tools with JSON inputs, execute returned tool requests in your code, send results back to Claude, and control repeated calls with a bounded agent loop.
Diagnose Claude API failures using HTTP status codes and error types. Identify authentication, request validation, and rate limit issues, apply appropriate fixes or bounded retries, and record verification results without exposing sensitive content.
Build a workflow that uses Figma webhooks, notifications sent when designs change, to queue Anima code generation. Handle approved sources, duplicate events, retries, and rate limits, with approval required before generated code is merged or deployed.
Upgrade Anima SDK integrations or move manual Figma exports into automated code generation. Compare generated files against a baseline, validate changes in staging, and document the package version, test results, and rollback reference.
Structure Anima SDK workflows with a reusable client, preset settings, source-aware caching, output normalization, and bounded retries. These patterns help teams make design-to-code generation more consistent while controlling cache scope and recovery behavior.
Generate React or HTML from a selected Figma frame using the Anima SDK. Explore styling options, inspect the generated files, and check formatting, types, dependencies, and appearance before bringing the code into an existing project.
Build a backend API that uses the Anima SDK to generate code from approved Figma files and nodes. The skill covers an Express wrapper and deployment options such as Vercel serverless functions and Cloud Run, with access and secret safeguards.