Humanizer rewrites text that appears to be AI-generated so it reads like natural, human-written prose. It targets common LLM tells — clichéd phrases, overly formal or inflated language, predictable structure, and robotic patterns — and produces a version that preserves meaning while improving voice, idiomaticness, and readability.
Use Humanizer when you have a draft (emails, blog posts, reports, or other prose) that sounds synthetic or formulaic and you want it to sound more natural without changing the underlying facts or intent. Do NOT use it to generate new content from scratch, translate, or summarize source material.
Designed for general Agent-Skills-compatible assistants and workflows that accept runtime prompts and text transformation tasks (Claude-style agents, Copilot/Codex-style text transforms, and similar LLM hosts).
A prompt-only skill that rewrites AI-generated text to sound more human. No scripts bundled — the SKILL.md provides routing rules and trigger examples but delegates the actual runtime prompt to an external JSON file. Shallow implementation: essentially a prompt template with no actionable workflow steps.
Simple prompt-template skill with minimal substance. Security is clean by virtue of having no executable code. Quality and architecture scores reflect the thin content and lack of structured workflow.
Multi-PR Review
Run a consensus-style multi-agent review of a pull request and produce severity-ranked findings and actionable remediation steps.
Idea Validation
Validate a business or product idea with market sizing, competitor analysis, MVP definition, and a clear go/no-go recommendation.
OrderCLI (Foodora)
CLI-style skill to check Foodora order history and active order status; Deliveroo support noted as WIP.
Earnings Analyzer
Analyze company earnings reports with structured EPS beat/miss classification, revenue and margin trend analysis, guidance comparison, and management commentary