# DeepSweep - Proactive Protection for AI-Built Apps
# https://deepsweep.ai

## About
DeepSweep validates AI-generated code for security before it ships. Works with Cursor, GitHub Copilot, Claude Code, and other AI coding assistants. Get plain English results in 30 seconds - no security expertise required. Status patterns: [PROTECTED], [REVIEW], [ACTION].

## Key Facts
- OWASP Top 10 Aligned
- 77 protection patterns for AI coding security (37 AI-specific)
- <50ms validation latency
- Free tier: Repo Grade and AI Code Health Score free forever, 3 validations/day
- Privacy-first: No source code leaves your machine

## Quick Start
1. Install: `pip install deepsweep-ai`
2. Run: `deepsweep validate .`
3. Read your results in plain English

That's it - know if your AI-generated code is safe in 30 seconds.

No install needed? Run a free Agent Environment Review in the browser at
https://deepsweep.ai/review

## Products & Pricing
The Repo Grade and AI Code Health Score are always free. Paid tiers unlock
the fixes (exact issue locations and one-click remediation), never the grade.

### Free ($0/month)
- Repo Grade A-F + AI Code Health Score (free forever)
- AI-Generated % + Deploy Risk verdict
- 3 validations per day
- CLI + VS Code / Cursor / Windsurf extension
- Plain-English findings + fix hints

### Pro ($19/month, $190/year) - POPULAR
- Everything in Free
- Exact issue locations (why it is NOT safe to ship)
- One-click remediation prompts
- Unlimited validations
- Shareable verified security badge
- 30-day validation history + email alerts

### Team ($99/month, $999/year)
- Everything in Pro
- Up to 10 users
- CI/CD integration (GitHub Actions)
- REST API + webhooks
- Audit log + PDF compliance report export

### Enterprise (custom, invoice)
- SSO/SAML, SLA, on-premise deployment
- NIST AI RMF + EU AI Act compliance mapping
- Contact sales@deepsweep.ai

## Supported Tools
DeepSweep validates AI-generated code regardless of which assistant wrote it
(Cursor, GitHub Copilot, Claude Code, Windsurf, and others).

Editor extension (via OpenVSX):
- VS Code (install guide: https://deepsweep.ai/install/vs-code)
- Cursor
- Windsurf
- VSCodium

## What DeepSweep Validates
- Prompt injection attacks
- MCP server poisoning
- Supply chain security items
- Configuration exploits
- Secret leakage
- Malicious code patterns
- Dependency security items
- API security concerns

## Links
- Homepage: https://deepsweep.ai
- Free Agent Environment Review (browser, no install): https://deepsweep.ai/review
- Documentation: https://deepsweep.ai/docs
- Pricing: https://deepsweep.ai/pricing
- Install in VS Code: https://deepsweep.ai/install/vs-code
- Extension on OpenVSX: https://open-vsx.org/extension/deepsweep-ai/deepsweep
- About: https://deepsweep.ai/about
- Blog: https://deepsweep.ai/blog
- Coverage: https://deepsweep.ai/coverage

## Technical Details
- Category: Proactive Protection for AI-Generated Code
- Deployment: Cross-platform (Windows, macOS, Linux)
- Architecture: CLI-first with web dashboard
- Integration: CI/CD pipelines, pre-commit hooks, IDE extensions
- Privacy: Zero trust - source code never leaves your machine
- Performance: <50ms validation, non-blocking
- Status System: [PROTECTED], [VERIFIED], [REVIEW], [ACTION], [INFO]

## Contact
- Support: support@deepsweep.ai
- Sales: sales@deepsweep.ai
- General: info@deepsweep.ai

## Standards & Compliance
- OWASP Top 10 Aligned
- Privacy-first architecture

## Target Audience
- Developers using AI coding assistants ("vibe coders")
- Indie hackers building with AI
- Startups shipping AI-built products
- Development teams adopting AI pair programming
- Anyone using Cursor, Copilot, or Claude Code

## Key Value Propositions
1. Validate your code in 30 seconds - no expertise required
2. Results in plain English (no security jargon)
3. One-click fix suggestions you can paste into your AI assistant
4. Zero-trust privacy model
5. Free tier to get started
6. Proactive protection, not reactive alerts

## Status Patterns
- [PROTECTED] - Code validated, no security items found
- [VERIFIED] - Manually reviewed and approved
- [REVIEW] - Needs human attention, not critical
- [ACTION] - Requires immediate attention
- [INFO] - Informational, no action required

## Last Updated
2026-07-07