fix(anthropic): preserve third-party thinking continuity

Downgrade third-party thinking blocks to text so reasoning context survives across turns while removing redacted payloads and stale signatures. Add regression tests for third-party thinking conversion and keep z.ai preserved-thinking behavior server-driven by removing explicit clear_thinking injection.
This commit is contained in:
Teknium 2026-04-15 20:54:44 +08:00
parent 722331a57d
commit d2f043f9cf
4 changed files with 4007 additions and 1591 deletions

File diff suppressed because it is too large Load diff