productivity▌

在开始联网操作前，先检查 CDP 模式可用性：

Set up a production-ready Playwright testing environment. Detect the framework, generate config, folder structure, example test, and CI workflow.

Create and manage Claude Code skills with auto-activation triggers, following Anthropic best practices and the 500-line rule. \n \n Two-hook architecture: UserPromptSubmit for proactive skill suggestions based on keywords and intent patterns, and Stop hook for gentle error-handling reminders after responses \n Three skill types with different enforcement levels: guardrail skills (block critical mistakes), domain skills (suggest best practices), and warning skills (advisory only) \n Configuration

AI-powered web search with real-time information and source citations via Perplexity models. \n \n Access five Perplexity models through a single OpenRouter API key, ranging from cost-effective Sonar to advanced Sonar Pro Search for complex multi-step analysis \n Ideal for finding current information beyond model knowledge cutoff, recent scientific literature, clinical trials, and domain-specific research with peer-reviewed sources \n Includes comprehensive query design guidance covering specifi

Diagnose and fix a Playwright test that fails or passes intermittently using a systematic taxonomy.

If you see unfamiliar placeholders or need to check which tools are connected, see CONNECTORS.md.

Measure what matters and make data-driven decisions.

Persistent knowledge capture for insights, decisions, and procedures that span multiple agent sessions. \n \n Store decisions with rationale, debugging lessons, repeatable workflows, and durable preferences as searchable memories \n Distinguish between new insights (use add ) and refinements to existing memories (use update ) to avoid duplication \n Design memories as atomic, standalone entries with clear titles that remain useful across future sessions \n Ideal for preserving incident learnings

Property-based fuzzer that compares Turso against SQLite to catch SQL correctness bugs. \n \n Generates random SQL statements and schemas, then executes them on both Turso and SQLite to detect mismatches in row sets, error handling, or schema state \n Supports deterministic reproduction via seed-based runs, configurable statement/table/column counts, and verbose output for debugging \n Includes continuous loop mode for extended fuzzing campaigns and Docker runner for CI with configurable timeout

Advanced techniques for maximizing LLM reliability, consistency, and output quality through systematic prompt design. \n \n Five core patterns: few-shot learning with 2-5 examples, chain-of-thought reasoning for multi-step logic, systematic optimization through A/B testing, reusable template systems with variables, and system prompts for persistent global behavior \n Progressive disclosure approach starts simple and adds complexity only when needed, with four escalation levels from direct instru