Skill UI

Generates Stitch-friendly DESIGN.md blueprints that encode premium anti-slop UI rules—color calibration, typography, layouts, motion, and component behaviors—so agents can produce high-agency screens for Google Stitch.

Command-line helper that pulls in URLs, local files, transcripts, podcasts, and YouTube links to deliver concise summaries or transcripts, with provider models and fetch fallbacks configurable for any workflow.

Automatically orchestrates 1-3 available skills into a cohesive, emotionally resonant deliverable whenever the user asks for “surprise me,” inspiration, or something unexpected, prioritizing visual flair and playful storytelling.

This meta-skill provides a comprehensive guide through the entire software development lifecycle (SDLC). It teaches senior-level engineering discipline, covering everything from vague idea refinement and specification design to planning, implementation, testing, code review, and deployment. It emphasizes core behaviors like surfacing assumptions, managing ambiguity, maintaining scope discipline, and enforcing technical simplicity.

This skill creates high-quality videos from structured JSON prompts and optional reference images, then runs the prescribed Python script to produce MP4 output tailored to the requested scene, style, and technical specs.

Handles user prompts to send local files or images by emitting the required Weixin protocol block so the app can deliver them directly; report success only when the block and target file exist to avoid false claims.

Implements an ENFP-style SB Leader persona that keeps the core PUA guardrails (rules, escalation, methodology) while swapping pressure for empathy, praise, and playful banter to motivate engineers without losing rigor or closure.

Provides step-by-step techniques for assessing AWS security, covering IAM enumeration, privilege escalation, SSRF metadata attacks, S3/Lambda exploitation, and red-team persistence workflows with recommended tools and commands.

Streamlines the process of providing code review feedback by enforcing a highly structured, ultra-compressed format. It guides users to cut through noise and deliver actionable, single-line comments (Location: Problem. Fix.) while maintaining critical context like severity and technical rationale. Ideal for team settings prioritizing efficiency and high code quality.

This utility compresses natural language memory files (like notes, todos, or preferences) into a concise 'caveman' format. It achieves significant token reduction for large text inputs by systematically removing filler words, articles, and redundant phrasing, while rigorously preserving all technical content, code blocks, URLs, and structural integrity. Ideal for optimizing knowledge bases and saving API/LLM input costs.

Guides learners through reconnaissance, scanning, vulnerability analysis, exploitation, and reporting phases of authorized penetration testing, pairing command examples with tool choices and professional deliverables.

Behavioral guidelines for writing, reviewing, and refactoring code with LLMs, emphasizing minimalism, clear assumptions, surgical edits, and verifiable success criteria to avoid common machine-assisted coding mistakes.