A stubbed status lookup returns an empty list every time. Callers read empty as clean, so a revoked or suspended credential sails straight through authorization.
What goes wrong
The function takes a credential and a set of status purposes, then returns an empty result no matter what you pass. Maybe it was a stub someone meant to finish. The danger lives entirely in how callers read that answer. An empty list of problems looks identical to "checked, found nothing," so a revoked credential and a healthy one produce the same output.
This is fail-open behavior dressed up as a normal return. There's no exception and no null. Nothing trips an error path. The call succeeds, monitoring stays green, and tests that only assert "no error" pass while the control does nothing.
The pattern
function lookupStatus(credential, purposes):
// TODO: query the status service
return [] // empty means no problems found
function authorize(credential):
problems = lookupStatus(credential, ["revoked", "suspended"])
if problems.isEmpty():
return ALLOW
return DENYfunction lookupStatus(credential, purposes):
result = statusService.query(credential, purposes)
if result == null:
throw StatusUnavailable() // missing is not clean
return result
function authorize(credential):
status = lookupStatus(credential, ["revoked", "suspended"])
if status.isRevoked() or status.isSuspended():
return DENY
return ALLOWGrounding and reachability were confirmed. The exploit-path gate would not conclude without the calling code, so the finding was held and an evidence request was raised instead of being reported as fact.
Why it matters to you
A control that fails open is worse than no control, because everyone downstream assumes it works. Your authorization flow trusts this answer to gate access. If the lookup is unfinished, unreachable, or misconfigured, you want the request to stop, not sail through. Decide what "unknown" means before an incident decides for you.
Check your own code
- Find every security check that can return empty or null, and confirm the caller treats that as a denial.
- Search for status or revocation stubs that return a hardcoded empty value.
- Write a test that feeds a known-revoked credential and asserts the request is refused.
- Make "status unavailable" a distinct outcome from "status clean."