Evaluation Provenance Disclosure: This interactive exhibit displays deterministic FIXTURE / SIMULATION playback of the canonical killer mission. It executes offline with zero live Google Calendar/Tasks mutations, zero AWS Bedrock/AgentCore API calls, and zero credentials required. Authentic historical live integration executions are preserved separately in the docs/ ledger (marked RECORDED_LIVE).
Canonical Intent (Mission 11111111):

“Get my family ready for tomorrow morning. We need to leave by 7:30.”

Alexa+ Spoken Response (7.2s):

“I noticed your departure moved to 8:00 AM on Google Calendar, so tomorrow morning is no longer ready for 7:30. I have added the backpack-packing task and confirmed it is on your list, and weather is checked.”

Current System-of-Record Truth:
DRIFTED
1
P-13.01

Seed Pre-State Verification

FIXTURE VERIFIED

Disposable demo calendar verified with 07:45 event; prep task absent.

Writes: 1 Read-back: VERIFIED
Inspect Evidence & Read-back Facts (2 properties)
prep_task_absentTrue
start_time07:45:00 UTC
2
P-13.02

Plan Intent & Observe Weather

FIXTURE OBSERVED

AWS planner compiled 3 actions (0 model authority). Weather observed with 0 writes.

Writes: 0 Read-back: NOT_APPLICABLE
Inspect Evidence & Read-back Facts (2 properties)
weathermild/rainy
model_authorityZERO
3
P-13.03

Create & Verify Prep Task

FIXTURE VERIFIED

Task 'Pack backpacks' created and independently verified via read-back.

Writes: 1 Read-back: VERIFIED
Inspect Evidence & Read-back Facts (2 properties)
task_titlePack backpacks
statusneedsAction
4
P-13.04

Pause at Calendar Approval Boundary

FIXTURE PAUSED_APPROVAL

Departure shift to 07:30 paused fail-closed. 0 writes performed without human grant.

Writes: 0 Read-back: NOT_APPLICABLE
Inspect Evidence & Read-back Facts (2 properties)
approval_requiredTrue
calendar_writes0
5
P-13.05

Approve, Execute & Verify Calendar Shift

FIXTURE APPROVED

Operator grant consumed. Departure updated to 07:30 and read back independently.

Writes: 1 Read-back: VERIFIED
Inspect Evidence & Read-back Facts (2 properties)
start_time07:30:00 UTC
writes_performed1
6
P-13.06

Produce READY Mission Receipt

FIXTURE VERIFIED

All predicates evaluated TRUE and FRESH. Mission promoted to READY with immutable hash.

Writes: 0 Read-back: NOT_APPLICABLE
Inspect Evidence & Read-back Facts (2 properties)
predicates_count3
receipt_hash5aa0ae701ac26800059f4618b2c642193d518a3fa3cce572c1d62d8488f92642

Receipt Hash: 5aa0ae701ac26800059f4618b2c642193d518a3fa3cce572c1d62d8488f92642

7
P-13.07

External Mutation & Drift Reconciliation

FIXTURE DRIFT_DETECTED

Departure moved to 08:00 externally. Mission reconciled READY -> DRIFTED. Historical receipt intact.

Writes: 0 Read-back: DRIFT_DETECTED
Inspect Evidence & Read-back Facts (2 properties)
observed_start08:00:00 UTC
predicate_truthFalse

Receipt Hash: 5aa0ae701ac26800059f4618b2c642193d518a3fa3cce572c1d62d8488f92642