skill-freshness

You can detect stale upstream data by running an automated freshness audit that examines articles, outputs, and memory state files to surface outdated dependencies before downstream skills consume them.

Upstream dependency freshness auditing is the process of identifying and surfacing stale upstream data across enabled skills to prevent downstream consumption of outdated articles or dead memory topics.

You monitor memory topics and outputs for staleness by examining explicit chains and implicit references against cadence-derived thresholds, generating per-consumer verdicts with severity bands like OK, WARN, STALE, or MISSING.

Yes, the freshness audit supports daily and weekly cadences by deriving cadence-based thresholds to compute severity bands and writing a daily article summary to report freshness degradation.

You generate automated reports for outdated dependencies by running the freshness audit, which writes an articles summary file and updates memory-topic state to ensure idempotent re-runs and track freshness degradation.

Automated freshness monitoring is limited to examining enabled skills, explicit chains, and implicit references in articles, outputs, and memory topics; it cannot assess data sources outside this configured scope.