Description
Right now test reliability suffers from flaky selectors and misleading comments, while documentation has gaps and contradictions — policy tests use broad locators that match multiple elements, test comments claim to remove limits but don't, CLAUDE.md is missing MCP resources, and rbac.md describes behavior that contradicts the implementation. Tests fail for the wrong reasons, and docs mislead contributors.
This epic tightens test reliability and doc accuracy — scoped selectors that target specific elements, comments that match what the code actually does, complete resource lists in CLAUDE.md, and rbac.md that documents real behavior. Tests become stable, and docs become trustworthy.
Description
Right now test reliability suffers from flaky selectors and misleading comments, while documentation has gaps and contradictions — policy tests use broad locators that match multiple elements, test comments claim to remove limits but don't,
CLAUDE.mdis missing MCP resources, andrbac.mddescribes behavior that contradicts the implementation. Tests fail for the wrong reasons, and docs mislead contributors.This epic tightens test reliability and doc accuracy — scoped selectors that target specific elements, comments that match what the code actually does, complete resource lists in CLAUDE.md, and rbac.md that documents real behavior. Tests become stable, and docs become trustworthy.