Problem
Large binary files or non-text assets can inflate the knowledge graph and waste tokens during indexing.
Solution
Add a --max-file-size flag (default: 1MB) to skip files exceeding the threshold during codegraph init. Accept human-readable values: 500kb, 2mb, 1gb.
Use case
CI pipelines indexing only source code, avoiding media files that would otherwise be included.
Problem
Large binary files or non-text assets can inflate the knowledge graph and waste tokens during indexing.
Solution
Add a
--max-file-sizeflag (default: 1MB) to skip files exceeding the threshold duringcodegraph init. Accept human-readable values:500kb,2mb,1gb.Use case
CI pipelines indexing only source code, avoiding media files that would otherwise be included.