docs(Agents): self-review-rules (#27043)

.agents/docs/self-review-rules.md

@@ -0,0 +1,26 @@
+# Self-review rules
+
+Project-specific rules for
+[/self-review](https://github.com/eai-org/agent-toolkit/blob/main/skills/self-review/SKILL.md).
+Add here rules about reviewing AzerothCore PRs before they are submitted. Generic improvements
+to the review process itself belong to the skill, not here.
+
+## Regression risk
+
+Automated test coverage is near zero, so this review is the main safety net against regressions.
+Every change gets the full review; go deeper the farther it can reach:
+
+- Shared core C++ (`src/server/game/`, `src/common/`, widely-included headers): one fix can
+  break unrelated features — examine how the changed code is used elsewhere, not just the
+  change itself.
+- Content scripts (`src/server/scripts/`) and modules: impact is mostly contained — worst case
+  is usually the one boss, spell, or module touched.
+- SQL: beyond the usual review, watch for the classic side effect — a DELETE/UPDATE whose
+  WHERE clause catches rows it shouldn't.
+
+## In-game testing
+
+PRs are expected to be tested in-game, which the reviewer cannot do. Never guess what the
+author already tested — ask them, and record the answer in the report. Then tell them what
+else to test — especially side effects they might not expect: a fix for X that also reaches Y
+means testing Y too. Invite them to ask questions if anything is unclear.

AGENTS.md

@@ -17,6 +17,7 @@ Read the matching doc(s) BEFORE starting the task:
   - Script work (under `src/server/scripts/`) → also `.agents/docs/cpp-scripts.md`
 - Creating or modifying SQL → `.agents/docs/sql-guidelines.md`
   - SmartAI work (`smart_scripts` data) → also `.agents/docs/cpp-scripts.md`
+- Self-reviewing a changeset before submission → `.agents/docs/self-review-rules.md`
 - Touching a subsystem that has a doc in `.agents/docs/systems/` → read that doc too
 - Capturing a lesson or adding/updating agent docs → `.agents/docs/README.md`