Knowledge Digest: Kill AI Slop + AI Cost Crisis
Signal 1: yetone/kill-ai-slop — AI Slop Detection & Removal Agent Skill
- GitHub: https://github.com/yetone/kill-ai-slop
- Website: https://killaislop.com
- Stars: 205 (created 2026-07-10, 1 day old)
- Language: TypeScript (Astro)
What it is: A two-part project: (1) a multilingual field guide (EN/ZH/JA/KO) cataloguing 23 visual/copy tells of AI-generated products; (2) an Agent Skill that scans web projects for these tells and strips them out.
23 detected tells include: indigo→violet gradients, gradient-clip headlines, warm “cozy” palettes, semantic color defaults, serif-italic emphasis, emoji everywhere, glassmorphism, glowing status dots, AI copywriting voice (“not just X — it’s Y”), badge & pill spam, pastel icon tiles.
Agent Skill design: Pure Node scanner (zero dependencies), SKILL.md-driven workflow with triage (confirm slop vs intentional), scan→report→fix approach. Uses npx skills add protocol.
Signal 2: AI Cost Crisis — Companies Curtailing Soaring AI Spend
- HN: https://news.ycombinator.com/item?id=48870035
- The Economist: Companies are scrambling to curtail soaring AI costs
- Points: 33 (up from 17 at collection)
Key insight: Enterprise AI spending moving from “unlimited experimentation” to “justified spending” with business cases and budgets. Parallels the cloud cost crisis of 2015-2018. Devs now responsible for P&L.
Impact: Expect AI FinOps tools to emerge. Open-source/local models gain relative advantage. Cost-per-task becomes competitive differentiator for agent ecosystems.