Command Entry Contracts¶
Each command has a narrow entry contract. The command should explain which governed repository contract it checked, where it read input, and what evidence the maintainer can trust.
Entry Contract Flow¶
flowchart LR
argv["command argv"]
root["workspace root<br/>resolution"]
input["governed input"]
validation["contract validation"]
output["maintainer evidence"]
argv --> root --> input --> validation --> output
Command-Specific Shape¶
| command | required contract | output standard |
|---|---|---|
audit-allowlist |
exception records use valid advisory ids, owners, links, reasons, and future expiries | diagnostics name the bad advisory |
deny-policy-deviations |
deviations have owner, reason, review link, and expiry tied to standards work | diagnostics name the bad deviation |
audit-ignore-args |
derived flags come only from reviewed allowlist records | output is suitable for CI consumption |
bench-compare |
benchmark snapshots compare against the governed baseline | strict mode fails on configured regressions |
Shared Entry Expectations¶
- Accept
--workspace-rootwhen repository root resolution is required. - Fail with the governed contract name, not only a parsing error.
- Keep successful validation output simple and automation-friendly.
- Write current-run benchmark evidence to documented generated locations.
- Do not add hidden reads or writes outside governed repository paths.
First Proof Check¶
Inspect crates/bijux-gnss-dev/src/main.rs,
crates/bijux-gnss-dev/docs/COMMANDS.md,
crates/bijux-gnss-dev/docs/WORKFLOWS.md, and
crates/bijux-gnss-dev/docs/GOVERNANCE_FILES.md.