Skill UI

Provides expert guidance for GRPO training with TRL, covering dataset prep, reward engineering, structured outputs, and multi-objective alignment when preference pairs are scarce.

Python framework for multi-objective optimization, offering NSGA-II/III, MOEA/D, benchmark suites, customizable operators, and visualization to help engineering teams explore Pareto fronts and make multi-criteria decisions.

This utility converts a single, high-level objective into a comprehensive, step-by-step build plan suitable for complex, multi-agent, or multi-session engineering projects. It automatically handles dependency mapping, detects parallel steps, and structures each step with self-contained context briefs, ensuring any new agent can execute it without prior knowledge. Ideal for large-scale refactoring or feature development spanning multiple PRs.

Blueprint turns a single objective into an executable, multi-step construction plan tailored for coding agents and multi-session teams, complete with context briefs, review gates, dependency awareness, and installable via `/blueprint <project> <objective>`.

Python framework for single- and multi-objective optimization with NSGA-II/III, MOEA/D, and SPEA2. Supports constraints, ZDT/DTLZ benchmarks, Pareto front analysis, and visualization for engineering design.

This skill integrates LangChain 1.0 and LangGraph 1.0 into standard OpenTelemetry (OTEL) backends (e.g., Jaeger, Honeycomb, Tempo). It solves critical observability gaps by ensuring correct prompt content capture, propagating spans across complex subgraphs, and providing five LLM-specific Service Level Objectives (SLOs) like latency and cost. Ideal for multi-tenant, compliance-sensitive, or existing OTEL stack environments where LangSmith is unsuitable.

A comprehensive OKR tracker dashboard designed to visualize team objectives and key results (KRs). It includes a quarter banner, multiple objective cards with progress bars, owner details, and a dedicated sidebar for at-a-glance KPIs, ideal for quarterly performance reviews and strategic planning.

Sends a board memo or strategy brief to multiple AI models (Claude, Codex, Gemini) for independent review, then reconciles consensus and divergences. Used for high-stakes, irreversible decisions like fundraises, pivots, layoffs, or regulatory commitments, offering an objective sanity check before moving forward.

This powerful skill pauses deep technical discussions to conduct a comprehensive, multi-dimensional reassessment. It analyzes conceptual drift, identifies unverified assumptions, and detects potential cognitive biases (e.g., confirmation bias, sunk cost fallacy). Use it when the conversation is becoming overly detailed, or when a strategic check-in is needed to ensure the focus remains on the core objective and the correct problem is being solved.

The Atlas Contract is a goal-integrity governance protocol designed for complex, multi-stage tasks such as backend development, API integration, or data persistence. It prevents 'goal drift' by enforcing critical checkpoints (Goal Contracts, Phase Checks, Final Audits). It is essential for high-risk, multi-part requests where maintaining strict alignment with the user's original objective is paramount.

Generates a complete, structured load and performance testing plan for any service. This skill guides the definition of test objectives, establishes measurable SLO thresholds (p99, error rate, throughput), and designs multi-scenario tests (Baseline, Stress, Spike, Soak). It is essential for teams needing to set clear pass/fail criteria, select appropriate tooling (k6, Locust), and integrate performance gates into CI/CD pipelines.

This skill applies the RICE framework (Reach, Impact, Confidence, Effort) to objectively score and rank product initiatives. It helps teams prioritize backlogs, plan quarters, and compare multiple competing ideas by providing a structured, data-driven ranking. The output identifies quick wins, high-risk moonshots, and critical data gaps for validation.