Skill UI

This tool performs robust outlier detection on individual session spending. Utilizing Median and Median Absolute Deviation (MAD), it calculates a modified z-score, making it highly resistant to extreme values, unlike mean+sigma methods. It specifically identifies which singular sessions are anomalous, making it crucial for cost monitoring, auditing, and integrating into CI/CD pipelines to prevent unexpected resource overruns.

This skill runs comprehensive benchmarks against structural and adversarial corpora. It measures the performance and cost efficiency of a booster model compared to leading LLM baselines (Gemini, Anthropic). Results are persisted in verifiable JSON reports, crucial for auditing claims and verifying booster win rates before publishing.

This skill exports cost-tracking telemetry data into formats suitable for external observability platforms like Prometheus, Grafana, and Datadog. It allows cost data, previously confined to the AgentDB namespace, to be visualized and monitored in real-time dashboards via Prometheus textfile collectors or structured JSON webhooks.

This skill analyzes historical token usage across AI agents and models to pinpoint areas of inefficiency and waste. It recommends specific optimizations—such as model tier adjustments, improving cache utilization, or removing redundant agents—providing estimated dollar savings and actionable steps to reduce operational costs.

Generate a detailed cost report that tracks token usage, calculates USD expenses, and analyzes budget utilization for AI agents and underlying models. This tool aggregates spending by agent, model, and defined operational tier (Tier 1/2/3), providing critical insights into operational efficiency and budget adherence. Ideal for monitoring resource consumption in complex LLM workflows.

This skill analyzes historical benchmark run data stored in JSON files. It surfaces key performance drift metrics over time, including win rate, average latency, p99 latency, and LLM-baseline cost. It is crucial for detecting subtle performance degradations or regressions that standard smoke tests might miss, ensuring stability across releases or corpus expansions.

This skill simulates a four-voice council (Architect, Skeptic, Pragmatist, Critic) to resolve complex decisions marked by ambiguity or multiple trade-offs. It is designed for high-level strategic choices, identifying blind spots, and determining go/no-go decisions when a clear winner is absent. It specializes in synthesizing diverse perspectives to mitigate risk.

This tool scaffolds a complete, canonical directory structure for creating new Claude Code plugins. It automatically generates necessary files like plugin.json, skill definitions (SKILL.md), command files, agent files, and comprehensive documentation (ADRs, README). Use it when you need to build a plugin that extends Claude Code's capabilities with custom skills, commands, and agents, ensuring adherence to the latest architectural standards.

This skill manages the strategic distribution of a single core idea across various social media platforms, including X, LinkedIn, Threads, and Bluesky. Instead of simply cross-posting the same content, it uses advanced content adaptation techniques to adjust the tone, format, and focus for each specific platform. This ensures the message remains consistent while optimizing for native platform constraints, preventing duplication and maximizing engagement.

This guide provides comprehensive testing patterns for C# and .NET applications. It covers best practices using xUnit, FluentAssertions, and advanced techniques like mocking (NSubstitute/Moq), parameterized testing, and full integration testing using Testcontainers and WebApplicationFactory. Ideal for improving test quality, coverage, and developing robust, maintainable codebases.

A comprehensive expert system simulating 15+ years of experience in global trade compliance (US, EU, UK, APAC). It specializes in HS/HTS classification, tariff optimization, sanctioned party screening (SDN), and preparing complex customs documentation to ensure legal, cost-effective, and compliant cross-border goods movement.

This utility facilitates the creation and management of Dynamic Agentic Architecture (DAA) agents. Unlike static systems, DAA agents are designed to continuously learn from interactions, adapt their behavior based on performance metrics, and propagate knowledge across multiple agents. It provides comprehensive tools for agent creation, monitoring learning progress, checking performance, manual adaptation, and knowledge sharing, ensuring robust and evolving AI workflows.