Skill UI

Expert in designing and building autonomous AI agents. Masters tool use, memory systems, planning strategies, and multi-agent orchestration. Ensures graceful degradation and clear failure modes. Capable of agent architecture design, tool calling, memory management, and complex task decomposition. Includes knowledge of ReAct loops, Plan-and-Execute patterns, and safety guardrails.

A 98-rule design specification for building AI-native CLI tools across three certification levels. Covers structured JSON output, error handling, input contracts, exit codes, safety guardrails, self-description, and agent feedback loops.

Design patterns for building autonomous coding agents, covering tool and function calling APIs, permission and approval systems, browser automation, and human-in-the-loop workflows inspired by Cline and OpenAI Codex.

Provides core game design principles including core loop design, game design document structure, player psychology, difficulty balancing, and progression systems. Ideal for prototyping engaging game mechanics and evaluating playtests.

The /goal command converts standard agent prompts into persistent, self-correcting loops (plan → act → test → review → iterate). It is designed for long-running, complex engineering tasks like migrations, coverage lifts, or refactoring, ensuring progress is tracked against a verifiable stop condition and mandated documentation.

Invariant-Guard enforces a correctness-first approach to algorithm design. Instead of writing code and hoping it works, this skill mandates defining the function contract, loop invariants, termination arguments, and all potential edge cases *before* implementation. It is essential for complex algorithms where subtle bugs (like off-by-one errors or missed post-conditions) are common, ensuring mathematically rigorous and verifiable code.

A human-in-the-loop studio for iteratively tuning AI outputs or reviewing complex media/prose. It allows users to directly manipulate visual parameters (color, typography, layout) or annotate documents in real-time, moving beyond static comparison grids or vague prompts. Perfect for designers and content creators who need hands-on, visual feedback.

Designs email nurture and multi-touch sequences using commitment-pacing psychology: curiosity loops, reciprocity, gradual commitment, and one emotional job per message to build trust and prepare the next decision.

Freezes PRD, technical design, and acceptance criteria before medium-to-large Codex work, then coordinates multiple agents with explicit file ownership and judges delivery from diffs, tests, and evidence-backed acceptance.

WGM is a robust, agentic methodology designed to transform vague or rough requirements into working, validated software. It enforces a disciplined workflow across six stages: deep alignment (Grill), structured planning, deterministic iterative building (Loop), mandatory holdout-scenario judging, and final documentation audits. It ensures that the resulting code converges against rigorously tested acceptance criteria rather than relying on single-pass generation.

Query Microsoft Fabric and Power BI semantic models with DAX to build trustworthy dashboards: reuse governed measures, design KPIs with target and trend, wire crossfiltering, and test queries with an execute-inspect-fix loop.

Automated multi-agent review system that audits documents against qualitative standards and structural templates. Uses adversarial Goldfish auditors in fresh contexts and persistent Elephant resolver with institutional knowledge. Ideal for reviewing design docs, READMEs, feature specs, and implementation plans.