Performance skills

This skill compresses conversation context across HOT, WARM, and COLD layers to reduce token usage and optimize memory during long discussions.

This skill enables autonomous self-improvement by safely updating configurations, skills, prompts, and memory to continuously optimize performance.

This skill analyzes code diffs and routes reviews to optimal models, delivering structured issues and cost-savings without sacrificing quality.

This skill provides a complete reference for building Miro Web SDK plugins and desktop apps, covering core APIs, authentication, events, and best practices.

This skill provides persistent, semantic memory for AI agents by storing and recalling important context across sessions with fast, local processing.

This skill provides local speech-to-text using Parakeet or Whisper backends for accurate English or multilingual fast transcription.

This skill helps you build and modify Go web services with ngamux, enabling fast route creation, middleware, and flexible responses.

This skill automatically tunes polymarket-executor parameters by analyzing performance metrics to improve future trading readiness and profitability.

This skill automatically routes tasks to the optimal model by complexity, saving cost while preserving performance for sub-agent tasks.

This skill reviews Q/kdb+ code for bugs, performance, and security, helping quant teams write correct, efficient kdb+ queries.

This skill unifies searches across Google, Wikipedia, Reddit, and RSS with Redis caching to deliver fast, comprehensive results.

This skill transforms the OpenClaw trader agent into an automated metals desk with real-time analysis, risk management, and executable trade orchestration.

This skill analyzes local log files to summarize statistics, detect duplicates, identify errors, and flag anomalies for quick investigation.

This skill helps automate ML competition evolution by learning from experiments and optimizing LB scores with memory-driven decisions.

This skill performs a comprehensive OpenClaw health check and delivers actionable fixes for configuration, sessions, models, security, and performance.

This skill helps you design, audit, and optimize OpenClaw deployments, from config analysis to health checks and infrastructure tuning.

This skill benchmarks your agent memory retrieval using a peer-review-grade suite, collects anonymized statistics, and submits results to research aggregations.

This skill enables fast, hybrid code and document search using BM25 and vector fusion, with chunking and MCP tools for easy access.

This skill enables fast local transcription using faster-whisper with GPU acceleration, providing word-level timestamps and flexible model options for accurate

This skill generates secure, parameterized SQL queries across multiple databases with validation, risk analysis, and audit-focused safeguards.

This skill generates a single high-performance X tweet for builders, delivering sharp, postable output with strict formatting and creator mode.