model-research

To keep Claude model selection guidance updated, run automated web research to check new model releases, pricing changes, and benchmark deltas against your existing repository configuration files. This ensures future coding workflow routing recommends the correct model tiers across sessions.

Comparing Claude model benchmarks and pricing automatically involves fetching official model documentation and performing web searches to detect context, max output, and benchmark deltas. The gathered data is then compared against your current model guide to identify what changed.

To apply researched Claude model updates to session and agent config files, execute the research command with the apply flag. This automatically applies diffs to the model guide, session context script, and batch task prompt configuration files.

Updating model selection guidance does not require manually editing markdown docs because the process automatically compares web research against the current models.md file. It optionally applies the necessary diffs directly to the configuration files.

You should run automated research for Claude model releases whenever new models drop or periodically to maintain accuracy. This keeps your coding workflow routing updated across sessions and agent assignments without manual checks.