§01
Sandbox
Test a function against sample input before submitting it for review.
§02
Test Results
| Test Case | Function | Status | Last Run |
|---|---|---|---|
| Standard invoice batch (20 files) | Receipt-to-Ledger Router | Passed | 2026-07-18 |
| Low-quality scan handling | Receipt-to-Ledger Router | Passed with warnings | 2026-07-18 |
| Duplicate detection across currencies | Invoice Duplicate Detector | Failed | 2026-07-17 |
| Non-standard indemnity clause set | Contract Clause Highlighter | Manual review required | 2026-07-16 |
§03
Automated Checks
Every test run is automatically screened for the following, in addition to whatever you test manually.
- Prompt injection vulnerabilities
- Sensitive-data leakage
- PII exposure
- Unsafe output
- Malware or harmful file handling
- Invalid API endpoints
- Excessive timeout or resource usage
- Missing output fields
- Broken input-output mapping
- Unsupported file types