Explorar skills

Configures end-to-end tracing for an LLM application using OpenTelemetry with LangSmith, Langfuse, or Helicone — span naming, metadata tagging, latency thresholds, and cost tracking.

This skill should be used when the user asks to "evaluate agent performance", "build test framework", "measure agent quality", "create evaluation rubrics", "implement LLM-as-judge", "compare model outputs", "mitigate evaluation bias", or mentions multi-dimensional evaluation, agent testing, quality gates, direct scoring, pairwise comparison, position bias, evaluation pipelines, or automated qualit

A Claude Skills collection providing ptrade‑dev strategy guardrails with lifecycle checks, API‑aware validation, and error‑pattern detection, plus structured parsing and webpage‑to‑Markdown tools—enabling LLMs to reliably analyze, generate, and debug PTrade/SimTradeLab strategies

Loop workflow for debugging a complex issue you've been stuck on for more than 30 minutes. Runs systematic hypothesis formation (debugging skill), feedback-loop mapping when the bug looks like a dynamic problem (systems-thinking), context-pathology check when it's an LLM/agent bug (context-degradation), blast-radius assessment before any fix attempt (risk-management), and uses test-driven-developm

Reviews an AI-generated response or LLM application output for factual risks, hallucination patterns, and confidence calibration issues.

Audits an AI application for unnecessary token spend and recommends prompt caching, model routing, and token reduction techniques to cut costs.

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools.

Guidance for designing, auditing, and improving context supplied to LLM agents, prompts, tools, RAG/MCP servers, and multi-agent systems. Use when the task involves context selection, prompt or agent architecture, tool descriptions, retrieval context, memory, compaction, autonomy, or evaluation of LLM behavior.

Diagnostic-loop workflow for improving an underperforming AI agent. Starts with baseline measurement using rubrics and LLM-as-judge scoring (agent-evaluation), then diagnoses architectural problems like task-model mismatch and pipeline bottlenecks (agent-project-development), redesigns the agent topology — supervisor, swarm, or specialized agents (multi-agent-patterns), adds the right memory frame

Designs an eval suite for an LLM agent or pipeline including success metrics, trajectory scoring, LLM-as-judge setup, and regression test cases.

Research-grounded 5-principle behavioral contract for production LLM coding agents — Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution, and Calibrated Communication. Each principle names its anti-patterns and failure modes with empirical backing (SWE-bench, FeatBench, DELEGATE-52, SE literature). Apply when implementing features, fixing bugs, refactoring, or any coding

Test and evaluate AI agents and LLM outputs using code-first evaluation framework with strong typing. Use when the user wants to: (1) Create evaluation datasets with test cases for AI agents, (2) Define evaluators (deterministic, LLM-as-Judge, custom, or span-based), (3) Run evaluations and generate reports, (4) Compare model performance across experiments, (5) Integrate evaluations with Pydantic