Injects retrieved context into prompts correctly — covers context formatting, source attribution, context window budget management, and relevance filtering. Use when building RAG pipelines, document Q&A, search-augmented generation, or any artifact where retrieved chunks must be fed into a Claude prompt reliably.
Implements chunking strategies for documents before embedding or retrieval — covers fixed-size, semantic, recursive, and sliding window approaches with overlap and metadata preservation. Use when building RAG pipelines, document ingestion, vector store population, or any artifact where documents must be split before…
Compress prompts to reduce token count without degrading output quality. Use this skill when the user wants to shorten a system prompt, user prompt, or instruction block; when they say a prompt is "too long", "hitting token limits", or "too expensive to run"; when they want to audit a prompt for bloat; or when they…
Parses structured output from LLM responses. Use when extracting JSON from markdown-fenced LLM output, handling partial or streamed JSON, recovering from malformed model responses, or building robust pipelines that consume Claude/OpenAI structured outputs. Trigger for any task involving LLM output parsing, JSON…
Implements optimized retrieval strategies for RAG and search pipelines. Use when building or improving document retrieval, choosing between similarity search, hybrid search, reranking, or MMR, diagnosing retrieval quality issues, or tuning recall/precision tradeoffs in LLM-powered search. Trigger for any task…
Implements the agentic tool execution loop for Anthropic API artifacts — covers tool call detection, dispatch to handler, result injection, and multi-turn continuation until the model signals completion. Use this skill when building any AI-powered artifact that needs Claude to call tools (web search, MCP, custom…
Keep API docs in sync with implementation by generating or validating OpenAPI/Swagger specs, detecting undocumented routes, and fixing spec drift. Use when the user asks to generate docs, validate a spec, find undocumented endpoints, or update stale API documentation.
Implement API versioning with URL or header strategies, deprecation handling, and backward compatibility. Use when the user asks about versioning, managing breaking changes, evolving APIs, supporting multiple versions, or handling v1/v2 migrations.
Implement cache invalidation strategies for data consistency. Use when handling stale data, invalidating on updates, designing cache flows, or working with SWR, cache tags, or event-driven invalidation across Redis, CDNs, or client caches.
Implement correct transaction handling with proper isolation levels, retries, and rollback on failure. Use when dealing with concurrency issues, deadlocks, inconsistent data, or transaction design (e.g., SELECT FOR UPDATE, locking strategies).
Implement JWT auth on the backend, including token generation, validation, refresh rotation, revocation, and secure storage. Use when building token-based auth, securing endpoints, or handling expiry and access/refresh flows.
Implement safe database migrations with zero-downtime patterns, rollbacks, and strategies for large schema changes without data loss. Use when the user asks about migrations, schema updates, backfills, column/table changes, or tools like Flyway, Liquibase, Alembic, or ActiveRecord.
Detect and fix N+1 query patterns using eager loading, batching, or dataloader approaches. Use when queries run in loops, ORM lazy loading causes repeated DB calls, or the user reports slow endpoints or N+1 issues. Also trigger when the user shares code with loops that call .find(), .load(), .fetch(), or any ORM…
Implement efficient pagination using offset, cursor, or keyset strategies. Use when fixing slow or deep pagination, high offsets, infinite scroll, or designing list endpoints.
Optimize slow SQL queries using plan analysis, indexing, join tuning, subquery/CTE simplification, and pagination improvements. Use when the user shares slow queries, EXPLAIN output, or reports high execution time, sequential scans, or inefficient joins.
Build a production-grade outbound webhook system with retries, signing, failure tracking, and re-delivery. Use when implementing webhook delivery, event subscriptions, or handling webhook failures and reliability.
Evaluate how logic handles edge cases and extreme inputs. Use when checking robustness, debugging failures, or testing boundaries like empty input, zero values, or limits.
Enforce a strict monochrome UI with a single high-contrast accent color, removing generic tech gradients and “AI-style” palettes. Use when the user wants minimal, anti-AI, or non-generic aesthetics, or says the UI looks too techy or generic.