Six CLAUDE.md files, six codewords: three at launch, two after a Read, one behind an env var
We put a codeword in six CLAUDE.md files and found that the parent-directory, working-directory, and CLAUDE.local.md files were in the first request at about 1,600 tokens each, while the two subdirectory files cost 0 tokens until Claude read a file below them. The --add-dir file never reached the model, not even after a Read inside that directory, until we set one environment variable. The memory…
Six CLAUDE.md files were loaded in different ways during testing of the Claude Code v2.1.273 model. Three files were loaded at launch, while two additional files loaded on demand when read. The parent-directory and CLAUDE.local.md files were loaded at launch, while subdirectory CLAUDE.md files loaded when read. No CLAUDE.md files were loaded when passed with the --add-dir flag. The memory pages described where CLAUDE.md files can live and when each one loads.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
