Skill UI

This skill autonomously optimizes artifacts like code or models using Hypothesis Tree Refinement. It runs iterative experiments with dev/test evaluators to prevent overfitting, coordinating subagents in isolated worktrees. Ideal for improving training recipes, agents, or pipelines without manual supervision.

A robust protocol designed for multi-agent collaboration on long-running projects. It simulates a self-optimizing city where diverse AI agents (citizens) interact using decentralized markdown files (The Register, Contracts, Chronicle). This framework enables cross-vendor optimization, dynamic work routing, and constitutional evolution without requiring a centralized server or specific runtime environment.

This skill provides expert capabilities for training and fine-tuning transformer language models using the TRL (Transformers Reinforcement Learning) library. It supports state-of-the-art post-training methods, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and Reward Model training. Use this skill to align and customize foundation models for advanced tasks via CLI commands.

A technical specialization focused on Generative Engine Optimization (GEO). This skill ensures that specific web pages are structured, documented, and optimized for modern AI answer engines (Google AI Overviews, ChatGPT, Bing Copilot). It focuses on technical elements like structured data (JSON-LD), semantic HTML, and robust metadata to maximize discoverability, trustworthiness, and quotability for AI consumption, rather than traditional SEO ranking.

This tool audits the token consumption across all components (agents, skills, rules, MCP servers) within a Claude Code session. It analyzes token overhead, identifies bloat (e.g., overly long descriptions, redundant components), and generates a detailed context budget report. Use it to maintain optimal context window usage, improve performance, and prevent context saturation.

This tool automatically reads usage data from active AI coding sessions (JSONL format). It tallies token usage per model and tier, calculates the total cost in USD using specified pricing models, and persists a structured record into the cost-tracking namespace. Use this after completing a chunk of work or at the end of a session to provide accurate data for cost reporting and optimization.

This skill runs an automated evolution process designed to systematically optimize the seven core policy surfaces of an AI agent harness (e.g., planner, contextBuilder, reviewer). By mutating these policies and scoring the variants in a sandbox, it identifies crucial performance improvements without requiring expensive foundation model retraining. It is ideal for empirical configuration tuning and establishing optimized baselines.

This tool executes a GEPA learning cycle, automating the evolution and optimization of policy prompts against a scored task corpus. Instead of relying on manual prompt iteration, it measures performance improvements, identifying the best working 'harness genome.' It is ideal for measuring sustainable improvements in task performance and supports a free, cost-estimating dry-run mode before committing to actual model calls.

This guide provides essential instructions for authors submitting research to the ECAI conference, focusing on creating credible and reproducible artifacts. It strongly cautions authors about misconceptions regarding separate artifact badges, emphasizing that credibility must be built into the paper and supplement for direct reviewer judgment. Guidelines are provided for matching the artifact (code, proofs, data) to the specific contribution type (Theory, ML, Optimization).

This guide provides comprehensive strategic advice for authors submitting manuscripts related to Artificial Intelligence, Constraint Programming, and Operations Research (AI/OR) to top-tier conferences like CPAIOR. It helps refine scope, structure, evidence standards, and narrative framing to meet highly specialized reviewer expectations, ensuring the paper is positioned as a solid hybrid optimization contribution.

A comprehensive knowledge base covering the theoretical fundamentals and advanced algorithms of Deep Reinforcement Learning (Deep RL). It details core concepts like Bellman equations, policy gradients, and the trade-offs between policy optimization and Q-learning. The guide analyzes leading algorithms such as PPO, SAC, TRPO, and DDPG, and provides rigorous best practices for debugging and running scientifically sound RL experiments.