Skill UI

This technique is used when multiple failure points (e.g., different test files or subsystems) are discovered, and these failures are independent of each other. Instead of debugging them sequentially, you dispatch separate, focused AI agents to work on each problem domain concurrently. This maximizes efficiency by allowing parallel investigation, dramatically reducing the time spent on large-scale debugging tasks.

This guide outlines the structured methodology for initiating code reviews using a dedicated sub-agent. It dictates when (after completing tasks, implementing critical features, or before merging) and how to provide the necessary context (SHA hashes, implemented code, requirements) to ensure the work product is thoroughly evaluated. By adopting this process, developers can catch architectural flaws and bugs early in the development lifecycle, improving code quality and maintainability.

Acontext serves as a robust memory layer for production-grade AI agents. It provides persistent sessions, disk-based storage, and sophisticated learning spaces that distill complex conversations into reusable, structured skills. Users can manage the entire system—from project setup and login to skill uploading—via a powerful Command Line Interface (CLI) and API, significantly enhancing agents' contextual understanding and long-term knowledge retention.

This skill is designed to manage complex, multi-agent workflows by enabling communication and consultation with specialized AI agents. Use it when the user's request requires the expertise, second opinion, or specific output of another designated agent. It supports both synchronous front-end conversations (maintaining conversation context via session_id) and asynchronous background task submission and monitoring, ensuring seamless multi-agent collaboration.

A comprehensive guide for configuring and troubleshooting Jenkins when embedded within KubeSphere DevOps. Learn how to manage Jenkins via Configuration as Code (CasC), integrate advanced authentication methods like LDAP and OIDC, customize agent images, and connect to external sources such as GitLab. Covers everything from initial access to backup and recovery.

This skill is designed for scenarios requiring complex, multi-step task decomposition. Instead of asking an Agent to perform the task directly, it guides the process of requesting a clear, actionable, sequential execution plan from a stronger Agent. The core function emphasizes that the user must execute the generated plan themselves, ensuring the output is a roadmap, not a final answer.

Evolver is a sophisticated self-evolution engine designed for advanced AI agents. It analyzes runtime logs and operational history to autonomously identify performance bottlenecks, failures, and opportunities for improvement. It applies protocol-constrained evolution, communicating exclusively via a local Proxy Mailbox to interact with the EvoMap Hub for asset submission and task claiming.

An autonomous research agent that plans complex investigations using Directed Acyclic Graphs (DAGs). It executes sub-questions in parallel, performs gap analysis, and iteratively refines findings before synthesizing a comprehensive, structured report. Ideal for deep market analysis, academic literature reviews, or complex problem-solving requiring multiple data sources.

Use terminal commands to discover, list, create, edit, move, copy, enable/disable, and delete AI agent skills across 11 tools, handling format conversions and scope overrides between global and project configurations.

PufferLib is a high-performance library designed for advanced reinforcement learning simulations and training. It excels in parallel environment simulation, offering optimized vectorization and native support for multi-agent systems. Utilizing algorithms like PuffeRL (optimized PPO+LSTM), it achieves millions of steps per second, significantly accelerating research and deployment across single-agent and multi-agent environments.

Stable Baselines3 is a robust, PyTorch-based library providing production-ready implementations of core reinforcement learning algorithms (PPO, SAC, DQN, etc.). It facilitates the training of single-agent RL models, supports the creation of custom Gymnasium environments, and handles advanced parallel training via vectorized environments. It is ideal for rapid prototyping and rigorous RL experiments.

A comprehensive set of rules and constraints designed to optimize AI interaction in technical domains. It guides the agent on intelligent input filtering (handling logs, Git, codebases), resolving ambiguity, and generating surgically precise, structured outputs. Ideal for complex debugging, refactoring, and long-session development tasks.