Skill UI

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.

Hierarchical Agent Memory (HAM) scopes Claude Code agent context by generating per-directory CLAUDE.md files, tracking token savings, and surfacing analytics via a dashboard so multi-folder projects avoid repeated full scans.

AgentMail provides a comprehensive REST API email infrastructure specifically designed for AI agents. Agents can provision dedicated inboxes, send and receive emails, handle external service verifications (e.g., GitHub signups), and manage real-time inbound webhooks. The system includes a 'karma' mechanism to ensure high domain reputation and prevent abuse, making it reliable for complex automation workflows.

This comprehensive guide outlines best practices for building reliable and robust AI agent harnesses. It covers optimizing action spaces, defining structured tools, managing observation formats, and implementing essential error recovery contracts. Use these principles to dramatically improve agent completion rates, manage context budgets effectively, and transition AI prototypes into stable, production-grade autonomous systems.

This protocol defines the structured communication rules for multiple specialized AI agents simulating C-suite roles (e.g., CFO, CTO, CRO). It dictates invocation syntax, response formatting, and enforces strict constraints (such as loop prevention, call depth limits, and isolation rules) to ensure coordinated, non-contradictory, and reliable cross-functional analysis for executive decision-making.

Aws Agentic Ai refers to an AI-driven agentic capability within AWS, aiming to automate complex workflows by orchestrating services and responding to evolving goals with minimal human intervention.

This skill implements the BDI (Belief-Desire-Intention) architecture for advanced cognitive agent modeling. It allows agents to reason about context by transforming external knowledge (like RDF triples) into internal mental states (beliefs, desires, intentions) and vice versa. Key features include goal-directed planning, explainable reasoning chains, and T2B2T (Triples-to-Beliefs-to-Triples) flow, making it ideal for multi-agent systems and augmenting LLMs with formal cognitive structures.

This pattern defines a robust, continuous loop structure for autonomous agents. It integrates crucial components like quality gates, evaluation harnesses, and explicit recovery controls. It guides complex workflows by defining decision paths (e.g., PR control, RFC decomposition) and includes strategies to mitigate common failure modes such as infinite loop churn or cost drift, ensuring measurable progress in autonomous systems.

This pattern addresses the fundamental limitation of context window size by treating the filesystem as a persistent, external memory. Agents can write large outputs, plans, or findings to files, offloading them from the active context. Instead of carrying all information, agents use targeted search and retrieval tools to load only the necessary context on demand, drastically reducing token bloat and improving performance over long, complex tasks.

This infrastructure provides a robust solution for running powerful AI agents in remote, sandboxed environments. It overcomes the limitations of local execution by ensuring unlimited concurrency, consistent environments, and collaborative features. Use this to build complex background coding agents, multi-client interfaces (e.g., Slack, Web), or scale agent workloads far beyond local machine constraints.

Analyzes Claude Code auto-memory, promotes stable learnings to CLAUDE.md/.claude/rules, extracts reusable skills, monitors MEMORY.md health, and provides review/promotion/extraction commands to keep project knowledge curated and actionable.