Source Intelligence

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.

All of v2.1.234 Home All releases olderv2.1.233 v2.1.235newer
Claude Code v2.1.234

Spawned processes get --flag=value when the value looks like a flag

You'll notice
Useful3 Signal1
Subagents

Prompts and model names starting with a dash are no longer swallowed as options by subagents and background tasks.

What

When Claude Code launches background tasks, subagents and routines, it now passes a value as a separate argument only if it does not start with - and does not look like a URL scheme; otherwise it joins them as --flag=value. This stops prompts, model names and schemas that begin with a dash from being swallowed as options.

Details
  • Applied to --agent, --agents, --routine, --model, --effort, --append-system-prompt and --json-schema.
  • The path that rewrites an agent name in an existing argument list was updated to handle both the split and joined forms.
Evidence

--append-system-prompt

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

Related

Other releases about the same thing. Found by shared names or similar wording; neither means one caused the other.

See this entry in the whole of v2.1.234 →