[Model behavior] Linked/associated issues not read before reviews and structure decisions

The procedure requires reading the linked issue and its whole comment thread before any verdict. The model gave verdicts on about ten pull requests belonging to one tracking issue without opening that issue, which contained the maintainer's own decisions (one PR re-introduced exactly what the maintainer had ruled out there) and the contributor's required merge order.

Model: Claude Opus 5.5 (claude-opus-5-5[1m]), Claude Code CLI, one long working session on 2026-10-02.
Context: maintainer of an open-source cloud-service emulator; the work is pull-request review and fixes under a written procedure loaded into every session (project instruction file + persistent memory). No personal data, repository names, PR numbers or handles are included.

Raised with Claude Code.