AMBIGUOUS PAYMENT RECOVERY SELF-TEST
Ambiguous Payment Recovery Kit
Reference: https://github.com/safal207/valta-pilot-lab

Purpose
-------
A timeout tells you nothing about the economic outcome. This self-test helps a
team decide whether a payment workflow can classify a post-timeout state as
zero, one, or unknown effects BEFORE an automated retry is permitted.

Answer each question with YES / NO / UNSURE. Any NO or UNSURE means an
automated retry on that path is not yet safe.

1. AUTHORIZATION BINDING
   Is authorization bound to the exact amount, recipient, target, and action?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

2. DURABLE EXECUTION OWNER
   Is one execution owner durably reserved before dispatch?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

3. FENCING
   Can stale workers be fenced after timeout or restart?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

4. ACCEPTED vs FINAL
   Is provider "accepted" kept separate from settlement finality?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

5. INDEPENDENT OBSERVATION
   Can the external economic effect be observed independently?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

6. UNKNOWN BLOCKS RETRY
   Does an unknown outcome block blind retry?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

7. REPLAYABLE EVIDENCE
   Can another process replay the evidence and reach the same verdict?
   [ ] YES  [ ] NO  [ ] UNSURE
   Notes:

Verdict vocabulary
------------------
VERIFIED            One matching economic effect is independently proven.
SAFE_TO_RETRY       Zero external effects and a pre-effect rejection are proven.
UNVERIFIED          The provider accepted, but the expected external effect is absent.
RECONCILE_REQUIRED  The effect may have happened or evidence disagrees; blind retry stays blocked.

Boundary statement
------------------
This material describes an evidence and verdict method. It does not process
payments, hold funds, or assert exactly-once execution for any system beyond a
declared, published test boundary.