Agent Skills · AI Agent
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
An automated local Skill evaluation workflow supporting single-Skill, multi-Skill, multi-model, and multi-framework comparisons.
A dispatch-based Agent efficiency system covering task checkpoints, token budgets, snapshots, self-repair, and evolution.
Runs three subagents in rounds to reach consensus on code, ideas, or papers.
Periodically checks inbox and group chats, separates notice and task items, executes due work, replies, and marks completion with Bearer auth and stale skipping.
Splits large work into CLI-agent tasks with bounded concurrency, model selection, output validation, and run reports.
Design, improve, and compress skill agents for the Claude Code and Codex ecosystem with gap analysis, overlap detection, Nexus integration, and governed self-evolution.
A skill-creation coach for agent scenarios that uses a four-layer structure, five-step workflow, reference-material feeding, and iteration practices to build testable, maintainable custom prompts.
An agent-native CLI marketplace for discovering, installing, and running professional software commands with JSON output.
Wraps the JJR IoT platform APIs to query device lists, property data, and images, with scheduled tasks, chart reports, and DingTalk message delivery.
Breaks down a scenario into automatable steps, searches, installs, and tests matching SkillHub skills, then produces a Feishu demo document with results.
Search and discover OpenClaw skills across ClawHub, the OpenClaw directory, LobeHub, GitHub, and community forums, with install tips and category references.
Breaks down complex problems into multi-step logical chains and outputs traceable reasoning processes with grounded conclusions.