Anti-Dark Pattern Linter
Automated CI/CD gatekeeper blocking deceptive, manipulative, and coercive front-end interface patterns before merge. Aligned with FTC 16 CFR Part 425 ("Click to Cancel") and EU Digital Services Act (DSA Article 25).
โก Quick Start โ Run in Your Pipeline
Execute directly in your terminal or continuous integration workflow:
# Run local scan across your project
npx anti-dark .
# Output GitHub Actions annotations to block pull requests
npx anti-dark . --format github
npx anti-dark .
# Output GitHub Actions annotations to block pull requests
npx anti-dark . --format github
๐งช In-Browser Interactive Rule Scanner
Test UI snippets against our deterministic dark-pattern heuristics:
๐ก๏ธ Deterministic Rules
- confirm-shaming: Flags emotionally manipulative opt-out copy.
- fake-urgency: Flags artificial countdowns and unverified scarcity.
- pre-selection: Flags pre-checked add-on boxes and marketing opt-ins.
- roach-motel: Flags mandatory exit surveys and asymmetric friction.
- hidden-costs: Flags surprise fees injected at final checkout steps.
- trick-questions: Flags double negatives and inverted toggles.
๐ค AI & Agent Integrations
Anti-Dark Pattern exposes standard Model Context Protocol (MCP) server endpoints and machine-readable instructions:
- LLM Guidelines: /llms.txt and /llms-full.txt
- MCP Server Card: /.well-known/mcp/server-card.json
- AI Catalog: /.well-known/ai-catalog.json
- XML Sitemap: /sitemap.xml