Scheduled context-to-action evidence
Loop Wake Reliability Auditor
A loop is not operational because it says success. Prove that it woke on time, used fresh context, ran once, and respected action guardrails.
WAKESchedule lag and overlap
CONTEXTSnapshot age at start
ACTIONApproval and key reuse
Audit a timestamped loop ledger
Load the six-run worked sample or paste an exported ledger, then audit.
Loop Wake Reliability Auditor / observed run evidence
AUDIT PENDING
0runs audited
0clear runs
0review runs
0minute envelope
0minute cadence deficit
Actual start-to-finish tape
RUNSTART LAGRUNTIMECONTEXT AGEOBSERVED EVIDENCE
Why success is insufficient
A completed action may still overlap another run, use stale context, repeat an action key, or bypass a required approval.
How to respond
Use a single-flight lock for overlap, retune cadence using observed occupancy, refresh context closer to action time, and enforce approval and idempotency at execution.
What is not measured
This local audit does not inspect hidden reasoning, contact live systems, or prove that a repeated key is harmful without external idempotency records.
Export unlocks after a verified audit.