Governance Model

Governance Model

mqgov uses the fail-closed mqclass classifier. Broker differences are reported through capabilities and unsupported operations return NOT_IMPLEMENTED rather than a fake success.

Tier	Meaning	Required Authorization
R0	Reads and previews: topic describe, group lag, message peek/tail, DLQ peek, ACL list, schema list/describe/check, fleet reads, dry-run	None, but audited
R1	Ordinary writes such as topic create, message produce, schema register for a new subject	--yes or interactive confirmation
R2	Elevated mutations such as topic alter, ACL grant, schema register for existing subject	--yes and non-empty --ticket
R3	Destructive/irreversible operations such as offset reset, topic purge/delete, schema delete, DLQ redrive/purge, broad ACL change, internal produce	R2 authorization plus the exact --allow-* flag

R3 Allow Flags

--allow-offset-reset, --allow-topic-purge, --allow-topic-delete, --allow-destructive-acl, --allow-internal-produce, --allow-schema-delete.

Protected contexts, protected topics, and internal/system topics raise the tier by one.

AI Agent Rules

1. Run mqgov capabilities -o json first because broker support differs.
2. Use -o json everywhere for integration paths.
3. Use --dry-run or --plan as the source of blast radius.
4. Stop and ask a human when a command requires ticket, allow flag, or high-risk confirmation.