chore: remove --max-warnings=0 from eslint commands

[dinwwwh]

Drops --max-warnings=0 from every eslint invocation so lint warnings no longer fail the check; only errors do.

Applies to the lint, lint:fix, and sync:sponsors scripts, the lint-staged pre-commit command, and the Claude Code post-edit hook. Same change is being applied to middleapi/orpc and middleapi/standardserver.

[pullfrog]

✅ No new issues found.

Reviewed changes

• package.json — dropped --max-warnings=0 from lint, lint:fix, and sync:sponsors, and from the lint-staged * entry. Warnings no longer fail these commands; only errors do.
• .claude/settings.json — dropped --max-warnings=0 from the PostToolUse eslint --fix hook, matching the package scripts.

I verified the sweep is complete on both sides of the diff: the base had exactly five --max-warnings=0 invocations (the four in package.json plus the Claude hook), all five are covered here, and the head tree has no remaining occurrences anywhere (including .zed/settings.json, dependabot.yml, and docs). CI (ci.yaml runs pnpm run lint) and the sponsors sync workflow are consistent with the new behavior, and eslint.config.js sets no warning thresholds that would conflict. The change is mechanical, internally consistent, and matches the stated intent — one deliberate trade-off to note is that warnings can now accumulate silently without an alternative gate, which is the point of the change.

  ^{｜ View workflow run ｜ Using DeepSeek Flash (default — pick a model for stronger reviews) ｜ 𝕏}