pii-scrubbing
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends the installation of official Braintrust and OpenAI SDKs from public registries (NPM, PyPI, Maven, RubyGems). These are standard dependencies for the skill's intended functionality and are maintained by trusted organizations.
- [COMMAND_EXECUTION]: Instructions include standard package management commands (pip install, npm install, go get, bundle install) and testing commands (go run, npx tsx, gradle run). These are used exclusively for environment setup and executing the provided local verification scripts.
- [DATA_EXFILTRATION]: The skill's primary purpose is to prevent data exposure by redacting PII. Analysis of the included networking code (e.g., in verify_local.py, smoke.rb, local-receiver.ts) confirms that all verification traffic is routed to local loopback addresses (127.0.0.1) using synthetic data. No unauthorized external network calls were detected.
- [SAFE]: The skill follows security best practices for telemetry redaction, using fixed-string replacement for sensitive keys rather than complex patterns that might be subject to bypass. It also provides clear guidance on how to verify the export path using the Braintrust CLI tool in a controlled manner.
Audit Metadata