Agent Skills · AI Agent
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
Uses Markdown state files and deterministic turn commands to drive LLM narration, stats, checks, NPCs, and logs.
A structured triphasic execution framework for AI agents that enforces execute, review, and advance loops with progress persistence, retry limits, and idle cutoffs.
A technical guide for authoring agent skills with TDD-style pressure tests, baseline failures, minimal documentation, and continuous loophole fixes.
Records AI agent errors, corrections, feature requests, and best practices in Markdown and promotes them into project memory or reusable skills.
Turn one-off stock tasks such as daily reports, candidate lists, risk checks, and order updates into recurring workflows, while flagging scheduling and delivery gaps.
A web information retrieval strategy for AI agents that branches across search, fetch, OpenCLI, and browser with explicit fallback reasons.
Guides scenario-based setup to generate a self-contained HTML workbench with drawer navigation, 6-8 modules, and localStorage persistence.
Build and manage CoPaw multi-agent systems with agent creation, trust setup, task routing, collaborative sessions, result aggregation, and workflow orchestration.
Let a Feishu group bot call an LLM with memory, web search, weather, and wiki skills, plus easy extension.
Turn natural-language phone tasks into TeddyMobile Vox custom outbound calls, guiding field completion and initiating the request.
Routes AI agent tasks to workers using capability profiles and dispatch rules from worker-roster.json while injecting org context.
PHOENIX v2.0 is a dual-loop evolution system for AI agents that combines skill prediction, dynamic composition, trust networks, memory management, and health monitoring.