Verify selected-source and assistant retrieval boundaries.
keryn knowledge compile <plan.json> --source <source.json> --out <new-directory> [--json]Options
Section titled “Options”Usage: keryn knowledge probe <drive-probe.json> --out <new-directory> [--json] keryn knowledge compile <plan.json> --source <source.json> --out <new-directory> [--json] keryn knowledge validate <suite.json> [--json] keryn knowledge run <suite.json> --out <new-directory> --suite-sha256 <digest> [--timeout <seconds>] [--json] keryn knowledge review <receipt.json> --suite <suite.json> --baseline <receipt.json> --baseline-suite <suite.json> --out <new-directory> Probe makes metadata-only Google Drive requests for explicit files using two configured token environment variables. Compile validates fresh source evidence, reviewed identities and synthetic fixtures offline. Source access does not prove assistant behavior. Run sends the reviewed assistant requests using the existing bounded AI security runner. Retain original suites and receipts for review. Probe exits: 0 expected source access observed; 1 unexpected source access; 3 inconclusive. Compile success is not a security pass.Examples
Section titled “Examples”# Success is not a security pass.keryn knowledge compile plan.json --source source.json --out out/suiteExit codes
Section titled “Exit codes”| Code | Meaning |
|---|---|
| 0 | pass |
| 1 | findings or a requested change gate |
| 2 | usage error or invalid input |
| 3 | incomplete or unknown completion |
| 4 | execution or normalization failure |