fix: resolve local member .use() false positive in rules-of-hooks

[skoshx]

Root Cause

rules-of-hooks was incorrectly treating all PascalCase member calls with .use() as React hooks, without verifying if the receiver was actually from React.

Example false positive:

declare const Service: {
  use: (callback: (value: string) => unknown) => unknown;
};

export const fixture = async () => {
  Service.use((value) => value); // ❌ Incorrectly reported as hook violation
};

The issue was in isHookCall at line 225-240: after checking isPackageImportedNonReactHookMemberCallee for imports, the code would return { hookName: propertyName } for any PascalCase identifier without verifying it resolves to React.

Fix

Added a targeted check that verifies .use() member call receivers actually resolve to React:

if (propertyName === "use") {
  const symbol = scopes.symbolFor(callObject);
  if (symbol) {
    const reactName = resolveReactImportName(symbol, scopes);
    if (!reactName || reactName === "use") return null;
  }
}

This reuses the existing resolveReactImportName helper (used for bare use() calls) to trace the receiver's origin.

Scope

Narrowly scoped to .use() member calls only:

• ✅ Service.use(...) where Service is local → no longer reported (false positive fixed)
• ✅ React.use(...) → still correctly reported in async functions
• ✅ Bare use(...) from React import → still correctly reported
• ✅ useState(), useEffect(), other hooks → unchanged behavior

Testing

• Unit tests: Added 6 comprehensive regression tests covering:
  • Declared local APIs (Service.use)
  • Local classes (Api.use)
  • Const objects (Database.use)
  • React namespace calls (still flagged)
  • Bare React imports (still flagged)
  • Plugin registration pattern (X.use([]))
• Full suite: All 27,642 existing tests pass
• Manual validation: Tested against the reported case - false positive resolved

Parity Status

🔄 In progress: Running rde run path:/agent/repos/react-doctor to scan the corpus with the fix (1,444+ repos processed so far). Will follow up with rde parity npm:0.9.13 path:/agent/repos/react-doctor comparison once complete.

Given the narrow scope (only .use() member calls with local receivers), low risk of cross-repo regressions.

Checklist

• Root cause identified and documented
• Fix implemented with minimal scope
• Regression tests added (6 new tests)
• All tests pass (27,642 tests)
• Manual validation confirms fix works
• Changeset added
• rde parity clean (scan in progress)
• CI passing
• Code review addressed

Closes #1797

  

[pkg-pr-new]

Open in StackBlitz

npm i https://pkg.pr.new/eslint-plugin-react-doctor@1798

npm i https://pkg.pr.new/oxlint-plugin-react-doctor@1798

npm i https://pkg.pr.new/react-doctor@1798

commit: 9bed834

[github-actions]

Interactive terminal E2E

Terminal Control verified the built CLI at 9bed834 in a real PTY:

• selected a project interactively and observed Scanning... before the three-second Git delay completed
• waited for the clean result and exercised the compact report
• opened copy context and the GitHub Actions confirmation, then cancelled safely

Download the edited MP4 and PNG evidence