zenlm/claude-sanitizer
1
0
Fork 0
mirror of https://github.com/zenlm/claude-sanitizer.git synced 2026-07-26 22:30:36 +00:00
Code Issues Packages Projects Releases Wiki Activity
6 Commits 1 Branch 0 Tags
main
Commit Graph
5 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Hanzo AI 092565df17 chore: symlink AGENTS.md + CLAUDE.md → LLM.md 2026-04-06 15:06:05 -07:00
Hanzo Dev 3c4297f08a Trigger CI: test runtime-generated secrets 2025-11-14 06:52:10 -08:00
Hanzo Dev d3a89857c7 Fix linting issues: remove unused imports and variables 2025-11-14 06:48:35 -08:00
Hanzo Dev f89b3a2b17 Add comprehensive test suite and CI 
- 50+ unit tests covering all sanitization scenarios
- Tests for API keys, crypto data, financial info, PII
- Tests for context preservation (file paths, function names)
- Edge case testing (empty strings, long text, multiple secrets)
- Real-world scenario tests (env files, config files, code)
- GitHub Actions CI workflow:
  * Multi-OS testing (Ubuntu, macOS, Windows)
  * Python 3.8-3.12 support
  * Lint with ruff
  * Security audit with bandit
  * Coverage reporting to Codecov
- pytest configuration with coverage
- Test dependencies in optional-dependencies
 2025-11-14 06:40:23 -08:00
Hanzo Dev e438f78960 Initial commit: AI-powered sanitization tool using Claude 
- Intelligent context-aware sanitization using Claude API
- Two-layer approach: regex pre-screening + AI analysis
- Supports both Haiku (cost-effective) and Sonnet (accurate)
- Detects API keys, crypto data, PII, financial data
- Preserves training value (file paths, function names, errors)
- Detailed reporting with severity levels
- Smart mode with keyword pre-screening to reduce costs
- Cost estimates: ~$0.0043 per 10K entries (Haiku)
 2025-11-14 06:08:58 -08:00