DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.
When the auto mode classifier transcript exceeds the model's context window, Claude Code now falls back to manual approval instead of aborting. In headless mode, the session is still aborted with a descriptive error.
Evidence
Context window handling (search for "Auto mode classifier transcript exceeded context window" and "Classifier transcript exceeded context window")
Strings lifted out of the shipped bundle, so the claim above can be checked against them.