Recheck the submitting principal and protect stored log output

This commit is contained in:
2026-11-18 10:10:06 +01:00
parent db2f5b407d
commit 0f15f91f65
16 changed files with 788 additions and 209 deletions
+11 -2
View File
@@ -53,8 +53,17 @@ Sensitive classifications are unioned across all scopes for the same target, so
choosing another principal/scope cannot declassify a source. ssh.service,
sshd.service, sshd@*.service instances, auditd.service and systemd-logind.service
are always sensitive, even if a configured sensitive list is empty. No membership
is seeded or changed by this patch. Stored result-reader APIs are unchanged;
this profile governs root execution authorization, not a new general data grant.
receives sensitive-log access by default.
A new grant for `query_logs` records the source classification on its job. The
immutable payload identifies the exact source and target. Stored log queries
and individual job reads enforce the reader’s current tenant permissions and
credential scopes. All log output requires `logs:read`. Sensitive sources and
historical jobs without a recorded classification also require
`logs:read_sensitive`. An operations reader can still see the job state and
numeric exit code. The result then carries `output_restricted: true` and omits
the protected output, evidence and free-text fields. Only explicitly selected
token scopes can convey the additional permission.
Root log jobs bind one exact unit and UTC interval, with nonnegative timestamps,
positive interval **at most exactly 24h** (no rounded-hour widening), 1–1000 lines,