Agent Skills · AI Agent
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
Curate persistent rules, transient context, and codebase structure for AI agents to improve output quality.
Calls MiniMax web search and image understanding tools via mcporter using a query or image URL.
Captures screenshots, uses a vision model to locate UI elements, and drives mouse and keyboard actions through win32api for CEF apps and visible windows.
Reads a target SKILL.md and its assets, scores it against a 100-point base and 30-point bonus checklist, and produces a structured review report with optional test cases.
Log errors, corrections, knowledge gaps, and feature requests in Markdown, then promote durable patterns to CLAUDE.md, AGENTS.md, and other agent memory files.
Analyzes surveillance images or videos for children's negative emotions such as crying, anger, and fear, then sends alerts and structured reports.
Let AI learn unfamiliar apps on demand using value assessment, usage watching, environment discovery, and memory decay.
Guides a four-stage interdisciplinary roundtable from needs diagnosis and guest casting to debate and actionable takeaways.
A maintenance-and-evolution meta-skill for AI agents that combines memory compaction, capability-gap analysis, prompt evolution, and skill distillation into reviewed changes.
Use input compression, structured reasoning, concise output, and context management to make AI answers shorter and save 30-50% of tokens.
Turns repeated workflows into triggerable, evaluable, and iterable AI agent skills via a four-layer architecture and six-step process, with SkillHub packaging support.
Code Immortality distills how people or topics think into a runnable agent skill via multi-source collection, mental model extraction, expression DNA, and self-evolution.