[umputun]

claude code emits You've hit your individual spend limit · run /usage-credits to raise it, or visit claude.ai/admin-settings/usage on team plans when a member's extra-usage budget runs out. That wording is not a superstring of any existing default: "individual spend" sits between "your" and "limit", so You've hit your limit cannot substring-match it.

result is the failure from #425 - patternError returns nil and the run dies with the raw claude exited with error: command wait: exit status 1 instead of a limit diagnostic.

add the string to both default lists, same shape as d70c894 (#361, session limit) and 3c461d3 (#336, org monthly limit): claude_error_patterns for a graceful exit when --wait is not set, claude_limit_patterns so wait_on_limit retries through the reset window.

on whether a spend cap belongs in the limit list: 3c461d3 already put the org monthly quota there, with the stated reason that wait_on_limit should retry through a monthly reset window. A team-plan individual spend cap is the same class, and --wait stays opt-in.

the phrase is long and specific, so it doesn't reintroduce the narrated-prose false positives fixed in #419/#321.

second commit is unrelated to the fix itself but was turned up while making it: CLAUDE.md claimed the claude_error_patterns defaults live in llms.txt, which carries the limit-pattern defaults only. The error list is mirrored in the README table instead. That line is the map a pattern addition follows, so the wrong pointer would send the next one to a file with nothing to update.

covered by two TestMatchPattern cases - one pins that the older entries genuinely miss this wording, one pins that the new string matches the real message including its · separator and trailing text.

Related to #425

[Copilot]

Pull request overview

Updates ralphex’s default Claude pattern detection so a newer Claude Code quota message (“You’ve hit your individual spend limit …”) is treated as a limit/error pattern, enabling correct diagnostics and optional --wait retry behavior.

Changes:

• Add You've hit your individual spend limit to embedded default claude_error_patterns and claude_limit_patterns.
• Add unit tests covering the miss with older patterns and the match with the new pattern.
• Fix documentation pointers in CLAUDE.md and update docs/defaults references (README.md, llms.txt).

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
README.md	Updates the documented default claude_error_patterns / claude_limit_patterns to include the new spend-limit string.
pkg/executor/executor_test.go	Adds TestMatchPattern cases for the individual spend limit wording (including the · separator).
pkg/config/values_test.go	Updates embedded-default expectations to include the new pattern in both Claude error and limit lists.
pkg/config/defaults/config	Extends the embedded default config’s claude_error_patterns and claude_limit_patterns values/comments.
llms.txt	Updates the documented default Claude limit patterns list to include the new spend-limit string.
CLAUDE.md	Corrects where defaults are documented (error patterns not in llms.txt) and clarifies limit-pattern default sources.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.