cocoscout

To reduce token waste during context loading, automatically rank and prepend only the most relevant local context sources before persona invocations. This ensures targeted data is supplied, minimizing irrelevant information and speeding up initial reasoning.

The best way to rank local context sources is to evaluate CocoGrove patterns, CocoContext files, environment snapshots, and prompt history using relevance scoring. This automatically identifies and assembles the most pertinent data before persona invocations to optimize reasoning.

Yes, CocoScout works with CocoPlus workflows by automatically evaluating and ranking environment snapshots, prompt history, and context files. It triggers automatic context loading when you initiate a build or flow run within your project.

Yes, you can use prompt history data to inform decisions as the system ensures only local, non-query data is used. It assembles relevant prompt history alongside local patterns and files to prepare context silently within a short runtime.