Feature request: Provide a text-based view option to allow use by agents

[MattFisher]

Scout is awesome and I'd love to be able to fully integrate it into agentic workflows.

Currently scout view provides a nice human-centred interface, but I can't see a way to view the results of a scan as plain text or json.

Claude used ls -la scans/scan_id=EP9AryWqWP786K7uBTxFi5/ and then python -c "import pandas as pd; df = pd.read_parquet('scans/scan_id=EP9AryWqWP786K7uBTxFi5/outcome_summary.parquet'); print(df.to_string())" which got the result, but isn't optimal.

[jjallaire]

You can call scan_results_df() to do basically what Claude did (you'd need to do any printing though):

https://meridianlabs-ai.github.io/inspect_scout/results.html#python-api