What problem does it solve?
This Skill prevents accidental leakage of personally identifiable information (PII) into logs, analytics, and LLM context by scanning source code and autonomously redacting sensitive values before they can be recorded or transmitted.
Core Features & Use Cases
- Automated PII scanning: Runs targeted grep-style phases over Dart and Python code to find prints, analytics, CoachContext construction, LLM prompts, and SharedPreferences usage.
- Pattern-based anonymization: Replaces detected IBANs, AVS/AHV numbers, emails, phones, NPAs, salaries, and employer names with standardized redacted tokens or ranges.
- Verification-first workflow: Requires fresh grep verification and flutter test results before committing fixes, with strict time and single-file iteration constraints.
- Use Case: Run a full repository scan to locate CRITICAL PII leaks in mobile app CoachContext and analytics calls, automatically redact one file at a time, verify the fix, and commit the change.
Quick Start
Run /autoresearch-privacy-guard to scan the repository, redact identified PII in one file at a time, verify with grep and flutter test, and commit verified fixes.