Agent Skills · AI Agent
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
Reduces AI conversation token usage by distilling intent, compressing input, and constraining output while preserving task quality.
Provides AI agents with a CLI-accessible Node.js runtime and helpers like snapshot, click, evaluate, and CDP for browser observation, UI interaction, and web automation.
Adds a local self-improvement and proactive maintenance workflow for AI agents, capturing corrections, preferences, and reflections in tiered memory.
A browser automation CLI for AI agents with navigation, extraction, screenshots, forms, proxy support, and human-agent collaboration.
Refine early ideas into reviewable design specs through one-question-at-a-time dialogue before implementation planning.
A CLI-based browser automation tool for AI agents, supporting snapshots, ref-based interaction, session reuse, screenshots, and tab control.
It uses Socratic follow-up questioning to clarify assumptions, expose gaps, and reach agreement before execution.
Provides memory management, nudge reminders, skill creation, and other modules to enable continuous self-evolution for AI Agents.
Generates bid proposals through a six-stage multi-agent pipeline with tender parsing, knowledge-base-first drafting, hard QA gates, five-dimension review, iteration, and three human checkpoints.
An engineering workflow for building, testing, and iterating skills with SKILL.md drafting, test prompts, baseline comparisons, and quantitative evaluation.
Adds a zero-intrusion visual preprocessing layer to text-only agents, using Qwen-VL to recognize screenshots, documents, and charts and return structured Markdown.
A file-based memory persistence solution for AI, solving issues like context loss between sessions and inability to learn from mistakes via a structured document system.