스킬
Ai Output Verifier
Make Your AI More Reliable. Enforce rigorous fact-checking, confidence scoring, and risk-based human gates to eliminate AI hallucinations. Implement rigorous verification, confidence scoring, source hierarchy, self-challenge, contradiction detection, and human-in-the-loop gates on any factual, numerical, recommendation, analytical, or high-stakes output. Activate for research summaries, claims, statistics, advice, plans, client-facing content, financial or legal language, code explanations, or whenever the user requests verified, audited, high-confidence, or reliable results. Use to reduce hallucinations and increase trustworthiness across all major agent platforms that support SKILL.md.
Anti Ai Website App Designer
Research-first UI and product design that never looks AI-generated. Use when designing apps, sites, screens, flows, design systems or any interface where the user wants original, category-aware, production-ready work instead of soft-blue templates, purple gradients or card grids. Triggers on design, UI, UX, app screen, landing page, mobile, web, dashboard, redesign, make it look premium, stop looking AI, award-winning design, native iOS Android, design system, design tokens. Runs research, three competing directions, build, subtraction, then measured verification with bundled scripts.
Git Commit Pr Writer
Turn your code changes into clean, professional Git commits and PRs. This AI skill analyzes your git diff, staged changes, branch summary, or change description and creates ready-to-use developer documentation without inventing changes. Generate clean, conventional commit messages and useful pull request descriptions from code changes. Write conventional commit messages and clear pull request descriptions from git diffs, staged changes, or branch summaries. Use when the user asks for a commit message, PR description, changelog entry, or has staged changes, a diff, or branch comparison to summarize.
