run biome check --linter-enabled=false on CI to check unsorted imports too

[dimaMachina]

No description provided.

[changeset-bot]

⚠️ No Changeset found

Latest commit: a41f03a

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
agents-docs	Ready	Preview	Comment	Nov 5, 2025 3:03pm
agents-manage-api	Ready	Preview	Comment	Nov 5, 2025 3:03pm
agents-manage-ui	Ready	Preview	Comment	Nov 5, 2025 3:03pm
agents-run-api	Ready	Preview	Comment	Nov 5, 2025 3:03pm

[claude]

Claude finished @dimaMachina's task —— View job

---

Code Review Complete

Todo List

• Read and understand repository structure and existing patterns
• Analyze changes in .github/workflows/ci.yml
• Analyze changes in package.json
• Research existing biome configuration and import sorting patterns
• Evaluate system-level architectural implications
• Provide comprehensive review feedback

---

Overall Assessment

This PR introduces a focused import organization check to the CI pipeline by adding biome check --linter-enabled=false as both a package script and CI step. The changes are minimal and well-scoped.

Detailed Analysis

✅ What Works Well

1. Minimal, Surgical Changes: The PR makes only the necessary changes - adding the script and CI step without touching unrelated code.

2. Consistent with Existing Patterns: The approach follows existing biome integration patterns already established in the codebase.

3. Strategic Use of --linter-enabled=false: By disabling linting rules, this focuses specifically on import organization, which aligns with the recent commit history showing import sorting work.

🔍 Areas for Consideration

1. Missing Context & Documentation

• High Priority: The PR has no description, making it unclear why this specific check is needed
• The --linter-enabled=false flag choice isn't explained - what problem does this solve that existing biome format doesn't?
• Consider adding a comment in the CI step explaining the purpose

2. Potential Redundancy with Existing Checks

• Medium Priority: The CI already runs pnpm format:check (line 94) which includes biome formatting
• Your biome.json shows "organizeImports": "on" in the assist section
• Need clarity on why both checks are needed - are they catching different issues?

3. CI Pipeline Positioning

• Low Priority: The new step runs after the main pnpm check but before the CI summary
• Consider whether this should be part of the parallel Turborepo checks for better performance
• Current positioning means it will always run even if earlier checks fail

4. Error Handling & Feedback

• Medium Priority: Unlike other checks, this doesn't contribute to the CI summary logic
• If this check fails, developers won't get clear guidance in the summary about how to fix it locally

🏗️ System Architecture Implications

1. Tooling Strategy: This continues the pattern of using Biome as the primary formatting/linting tool, which is consistent with the existing architecture.

2. CI Performance: Adding a sequential step rather than parallelizing with Turborepo may impact CI times, though the impact should be minimal for this check.

3. Developer Experience: The biome:check script provides local reproducibility, which is good for DX.

📝 Specific Recommendations

1. Add PR Description (Critical)

Explain:
- What import issues this catches that existing checks miss
- Why --linter-enabled=false is needed
- Examples of problems this prevents

2. Consider CI Integration Options

# Option A: Add to existing parallel checks in turbo.json
# Option B: Add conditional logic to CI summary
# Option C: Document why sequential execution is preferred

3. Add Inline Documentation

- name: Biome Check (Import Organization)
  run: pnpm biome:check
  # Explain: Checks import organization without running linter rules
  # This complements format:check by catching import sorting issues

🎯 Questions for Clarification

1. What specific import organization issues were discovered that prompted this change?
2. Why use --linter-enabled=false instead of configuring specific rules?
3. Should this check block CI, or would it be better as a warning?
4. Is there a plan to eventually integrate this into the main pnpm check pipeline?

🔒 Security & Best Practices

• ✅ No security concerns identified
• ✅ Uses existing biome version and configuration
• ✅ No new external dependencies
• ✅ Follows existing CI patterns

Verdict

This is a well-executed, minimal change that addresses a specific need. The main gap is missing context - once the PR description explains the "why" and "what problems this solves," this would be ready to merge.

The change follows good engineering practices: small scope, consistent patterns, and provides local reproducibility. Consider the CI integration suggestions for optimal developer experience.

[robert-inkeep]

@claude format:check to me sounds like it shouldn’t ever write

[claude]

Claude finished @robert-inkeep's task —— View job

---

Clarification on format:check naming convention

Todo List

• Read package.json to understand current scripts
• Verify that format:check and biome:check are read-only
• Provide clarification about the naming and behavior

---

You're absolutely right, @robert-inkeep! The naming convention for "check" commands implies they should be read-only and not modify files. Looking at the PR changes:

Current Behavior Analysis

The PR modifies these scripts:

// Before:
"format": "biome format --write",
"format:check": "biome format",

// After (this PR):
"format": "biome format --write && biome check --linter-enabled=false --write",
"format:check": "biome format && biome check --linter-enabled=false",

The Issue

You're correct - the format:check command should not write files. However, looking at the Biome CLI behavior:

1. biome format (without --write) - Checks formatting but doesn't modify files ✅
2. biome check --linter-enabled=false (without --write) - Checks import organization but doesn't modify files ✅

Both commands in format:check are indeed read-only as they don't include the --write flag. They only report issues without fixing them.

Script Behavior Summary

Script	Modifies Files?	Purpose
format	✅ Yes	Fixes formatting AND import organization
format:check	❌ No	Only checks for issues without fixing

What This PR Does

The PR adds import organization checking to both scripts:

• format now also fixes import organization (with --write)
• format:check now also checks import organization (without --write)

The --linter-enabled=false flag ensures only the "assist" features (like organizeImports) run, without running the full linter rules.

Conclusion

The naming is appropriate - format:check remains read-only as it should be. Both biome format and biome check --linter-enabled=false without the --write flag only report issues without modifying any files.

---