Know which augmentation modules can handle each weak sample.
Splits weak VCN samples into label-eligible subsets for mining and anomaly generation, with an auditable routing report.
When to use it
Use immediately after gap analysis in each VCN AOI SDA iteration to route weak samples to augmentation modules.
Give it a gap-analysis parquet and, optionally, a source-pool CSV; it writes two routed parquets, a per-label summary, and a routing report.
What you provide
Requires Python to run the routing recipe through Bash.
Requires pandas for the routing recipe's data loading, filtering, and parquet output.
This skill
~/.claude/projects/
Reads your agent session transcript
/tmp/routing-hook-debug.log
Appends hook debug logs