0.1.8 RC1: Run the fixed GLM-5.2 Chat regression batch #289

Parent

#257 (part of #248)

Outcome

Run and publish the immutable fixed maintained-Provider regression batch for Ollama Cloud GLM-5.2 over Chat Completions on the exact RC candidate using the shipped Beta6 Chat scope.

Acceptance criteria

• The GLM Chat profile records 10/10 clean fixed cases without selective replacement.
• Progressive streaming, multi-turn text, standard tools, default reasoning, classified errors, and the real read/apply_patch/verify loop are represented.
• Exact Provider/model identity remains unchanged; no hidden Responses, Official fallback, or cross-Provider tool execution counts as success.
• Failed, degraded, superseded, and external-availability attempts remain visible.
• Evidence is bounded, sanitized, candidate-bound, and cleaned on every exit path.
• The result is maintained-Provider release evidence, not a runtime model allowlist.

Blocked by

• 0.1.8 Beta7 gate: Publish the stability and release-evidence candidate #319