Skill UI

Explains the altitude-and-horizon mental model that differentiates PM and Director thinking, helping leaders diagnose friction zones, create cascading context, and focus on system-level strategy when scope or direction is unclear.

Helps product managers diagnose whether AI workflows rely on indiscriminate context stuffing or deliberate context engineering, then apply bounded domains, retrieval, and Research→Plan→Reset→Implement practices to improve reliability and output quality.

This skill provides comprehensive techniques for optimizing prompts used with any Large Language Model (LLM). It teaches proven methodologies—including Role definition (R), Context setting (S), Constraints (C), Instructions (I), and Templates (T)—to transform vague inputs into precision-engineered instructions. Users can reliably achieve structured JSON output, implement Chain-of-Thought reasoning, reduce hallucinations, and minimize token usage, ensuring consistent and high-quality results across different AI platforms.

Guides product managers through choosing the right prioritization approach (RICE, ICE, Value/Effort, etc.) by asking adaptive questions about stage, team context, stakeholder needs, and decision cadence, then delivers tailored implementation guidance.

This comprehensive guide outlines best practices for building reliable and robust AI agent harnesses. It covers optimizing action spaces, defining structured tools, managing observation formats, and implementing essential error recovery contracts. Use these principles to dramatically improve agent completion rates, manage context budgets effectively, and transition AI prototypes into stable, production-grade autonomous systems.

A guide for creating and maintaining the canonical `AGENTS.md` file. This skill enforces minimal, high-signal documentation, ensuring agents receive concise instructions regarding package management, file-scoped commands, and project conventions, while strictly omitting filler or obvious context.

This skill establishes a rigorous framework for building deep architectural context by performing ultra-granular, line-by-line analysis of code. It applies methodologies like First Principles and 5 Whys to maintain a stable, evolving mental model of the entire system, making it essential for pre-vulnerability security auditing, threat modeling, and complex architecture reviews.

This skill implements the BDI (Belief-Desire-Intention) architecture for advanced cognitive agent modeling. It allows agents to reason about context by transforming external knowledge (like RDF triples) into internal mental states (beliefs, desires, intentions) and vice versa. Key features include goal-directed planning, explainable reasoning chains, and T2B2T (Triples-to-Beliefs-to-Triples) flow, making it ideal for multi-agent systems and augmenting LLMs with formal cognitive structures.

This advanced engine acts as the memory layer for C-suite advisory sessions. It loads, manages, and enriches detailed company context from local files, automatically detecting context staleness. Crucially, it enforces strict anonymization protocols before any external API calls, ensuring sensitive data (like revenue or names) is protected while providing deep, specific insights for high-level business strategy.

Transforms vague UI ideas into structured, Stitch-optimized prompts by adding platform context, UI/UX keywords, design system cues, and detailed page structure guidance so generated interfaces stay consistent and polished.

This skill provides a comprehensive framework for the continuous self-improvement of makepad-skills. It outlines automatic skill triggering via hooks, context-aware skill routing, and a structured process for capturing new patterns, error solutions, and best practices. It ensures the entire skill library remains robust and up-to-date throughout the development lifecycle.

This pattern addresses the fundamental limitation of context window size by treating the filesystem as a persistent, external memory. Agents can write large outputs, plans, or findings to files, offloading them from the active context. Instead of carrying all information, agents use targeted search and retrieval tools to load only the necessary context on demand, drastically reducing token bloat and improving performance over long, complex tasks.