Back to Details

pr-review-fixer

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

PR Review Fixer

PR评审修复工具

Fetch unresolved PR comments, validate each issue, create a fix plan, implement fixes, and verify CI checks pass (tests, lint, build).
获取未解决的PR评论，验证每个问题，制定修复计划，实施修复并验证CI检查通过（测试、Lint、构建）。

Workflow

工作流

1. Fetch PR Comments

1. 获取PR评论

bash
undefined
bash
undefined

Get PR info

Get PR info

PR_NUM=$(gh pr view --json number --jq '.number')
PR_NUM=$(gh pr view --json number --jq '.number')

Get all PR comments via GraphQL (code-level AND PR-level)

Get all PR comments via GraphQL (code-level AND PR-level)

gh api graphql -f query=' query($owner: String!, $repo: String!, $pr: Int!) { repository(owner: $owner, name: $repo) { pullRequest(number: $pr) { # Code-level review comments (file/line specific) reviewThreads(first: 100) { nodes { id isResolved comments(first: 50) { nodes { id body author { login } path line } } } } # PR-level review comments (top-level review body) reviews(first: 50) { nodes { id body state author { login } } } # PR-level issue comments (general discussion) comments(first: 100) { nodes { id body author { login } } } } } } ' -f owner=OWNER -f repo=REPO -F pr=$PR_NUM
undefined
gh api graphql -f query=' query($owner: String!, $repo: String!, $pr: Int!) { repository(owner: $owner, name: $repo) { pullRequest(number: $pr) { # Code-level review comments (file/line specific) reviewThreads(first: 100) { nodes { id isResolved comments(first: 50) { nodes { id body author { login } path line } } } } # PR-level review comments (top-level review body) reviews(first: 50) { nodes { id body state author { login } } } # PR-level issue comments (general discussion) comments(first: 100) { nodes { id body author { login } } } } } } ' -f owner=OWNER -f repo=REPO -F pr=$PR_NUM
undefined

2. Filter Comments

2. 过滤评论

Code-level comments (reviewThreads):
  1. Exclude resolved threads: Filter out threads where 
    isResolved: true
  2. claude[bot] handling: For 
    claude[bot]
    comments, keep only the last comment per thread
  3. Group by file/location: Organize by path and line number
PR-level review comments (reviews):
  1. Exclude empty bodies: Skip reviews with empty or whitespace-only body
  2. Exclude approval-only: Skip reviews with state 
    APPROVED
    and no actionable feedback
  3. claude[bot] handling: For 
    claude[bot]
    reviews, keep only the most recent one
PR-level issue comments (comments):
  1. Exclude bot noise: Skip automated comments (CI bots, etc.) unless actionable
  2. claude[bot] handling: Keep only the last 
    claude[bot]
    comment
  3. Identify actionable items: Look for requested changes, questions, or suggestions
代码级评论（reviewThreads）：
  1. 排除已解决的线程：过滤掉
    isResolved: true
    的线程
  2. claude[bot]评论处理：对于
    claude[bot]
    的评论，每个线程仅保留最后一条评论
  3. 按文件/位置分组：按路径和行号整理评论
PR级评审评论（reviews）：
  1. 排除空内容：跳过内容为空或仅含空白字符的评审
  2. 排除仅批准的评审：跳过状态为
    APPROVED
    且无可执行反馈的评审
  3. claude[bot]评论处理：对于
    claude[bot]
    的评审，仅保留最新的一条
PR级问题评论（comments）：
  1. 排除机器人噪声评论：跳过自动化评论（如CI机器人评论），除非包含可执行内容
  2. claude[bot]评论处理：仅保留最后一条
    claude[bot]
    评论
  3. 识别可执行项：查找要求修改、问题或建议内容

3. Determine Working Location

3. 确定工作位置

Review reports are posted as PR comments (not committed). Task files are local working files only.
Working directory: Use a temp directory or 
.claude/reviews/PR-[number]/
for local task files during the session. These files are not committed.
Iteration tracking: Check existing PR comments by 
claude[bot]
with "PR Review Overview" in the body. Count those to determine the current iteration number N.
评审报告将作为PR评论发布（不提交到代码库）。任务文件仅保存在本地工作目录中。
工作目录：使用临时目录或
.claude/reviews/PR-[number]/
作为会话期间的本地任务文件存储目录。这些文件不会被提交。
迭代跟踪：检查PR中由
claude[bot]
发布的、正文包含"PR Review Overview"的评论，统计这些评论的数量以确定当前迭代次数N。

4. Validate Issues

4. 验证问题

Code-level comments:
  1. Read the referenced code at path:line
  2. Evaluate: Is issue still present? Is suggestion correct? Does it align with project conventions?
  3. Mark as valid or invalid with brief rationale
PR-level comments:
  1. Parse the comment for actionable items (suggestions, questions, requested changes)
  2. Check if the feedback applies to current PR state (changes may have been made since)
  3. Evaluate: Is the request reasonable? Does it align with project goals?
  4. Mark as valid or invalid with brief rationale
  5. Skip pure acknowledgments, thanks, or informational comments without action items
代码级评论：
  1. 读取路径:行号对应的代码
  2. 评估：问题是否仍然存在？建议是否正确？是否符合项目规范？
  3. 标记为有效或无效，并附上简短理由
PR级评论：
  1. 解析评论中的可执行项（建议、问题、修改要求）
  2. 检查反馈是否适用于当前PR状态（自评论发布以来可能已进行过修改）
  3. 评估：请求是否合理？是否符合项目目标？
  4. 标记为有效或无效，并附上简短理由
  5. 跳过纯确认、感谢或无执行项的信息性评论

5. Create Review Overview

5. 创建评审概述

Prepare the review overview content (this will be posted as a PR comment, not committed):
markdown
undefined
准备评审概述内容（将作为PR评论发布，不提交到代码库）：
markdown
undefined

PR Review Overview - Iteration [N]

PR Review Overview - Iteration [N]

PR: #[number] | Branch: [name] | Date: [YYYY-MM-DD]
PR: #[number] | Branch: [name] | Date: [YYYY-MM-DD]

Valid Issues

Valid Issues

Code-Level Issues

Code-Level Issues

Issue 1: [title]

Issue 1: [title]

  • File: 
    path:line
  • Reviewer: @user
  • Comment: [quoted]
  • Validation: [rationale]
  • File: 
    path:line
  • Reviewer: @user
  • Comment: [quoted]
  • Validation: [rationale]

PR-Level Issues

PR-Level Issues

Issue 2: [title]

Issue 2: [title]

  • Type: review comment | discussion comment
  • Reviewer: @user
  • Comment: [quoted]
  • Validation: [rationale]
  • Type: review comment | discussion comment
  • Reviewer: @user
  • Comment: [quoted]
  • Validation: [rationale]

Invalid/Skipped Issues

Invalid/Skipped Issues

Issue A: [title]

Issue A: [title]

  • Location: 
    path:line
    or PR-level
  • Reviewer: @user
  • Comment: [quoted]
  • Reason: [why invalid]
undefined
  • Location: 
    path:line
    or PR-level
  • Reviewer: @user
  • Comment: [quoted]
  • Reason: [why invalid]
undefined

6. Create Task List

6. 创建任务列表

Use rune to create 
review-fixes-[N].md
:
bash
rune create ${OUTPUT_DIR}/review-fixes-${N}.md \
  --title "PR Review Fixes - Iteration ${N}" \
  --reference ${OUTPUT_DIR}/review-overview-${N}.md
使用rune创建
review-fixes-[N].md
：
bash
rune create ${OUTPUT_DIR}/review-fixes-${N}.md \
  --title "PR Review Fixes - Iteration ${N}" \
  --reference ${OUTPUT_DIR}/review-overview-${N}.md

Add tasks via batch for efficiency

Add tasks via batch for efficiency

rune batch ${OUTPUT_DIR}/review-fixes-${N}.md --input '{ "file": "review-fixes-'${N}'.md", "operations": [ {"type": "add", "title": "Fix: [issue 1]"}, {"type": "add", "title": "Fix: [issue 2]"} ] }'
undefined
rune batch ${OUTPUT_DIR}/review-fixes-${N}.md --input '{ "file": "review-fixes-'${N}'.md", "operations": [ {"type": "add", "title": "Fix: [issue 1]"}, {"type": "add", "title": "Fix: [issue 2]"} ] }'
undefined

7. Fix Issues

7. 修复问题

Loop through tasks:
  1. rune next [file]
    - get next task
  2. rune progress [file] [id]
    - mark in-progress
  3. Implement the fix
  4. rune complete [file] [id]
    - mark complete
  5. Repeat until done
循环处理任务：
  1. rune next [file]
    - 获取下一个任务
  2. rune progress [file] [id]
    - 标记为进行中
  3. 实施修复
  4. rune complete [file] [id]
    - 标记为完成
  5. 重复直到所有任务完成

8. Check CI Status

8. 检查CI状态

After fixing review comments, verify CI checks:
bash
undefined
修复评审评论后，验证CI检查状态：
bash
undefined

Get check status for the PR

Get check status for the PR

gh pr checks --json name,state,conclusion
gh pr checks --json name,state,conclusion

For failed checks, get details

For failed checks, get details

gh run view [RUN_ID] --log-failed

**Check types to handle:**

1. **Test failures**: Parse test output, identify failing tests, fix code or tests
2. **Lint errors**: Run linter locally, fix reported issues
3. **Type errors**: Run type checker, fix type mismatches
4. **Build failures**: Check build logs, fix compilation issues
gh run view [RUN_ID] --log-failed

**需要处理的检查类型：**

1. **测试失败**：解析测试输出，识别失败的测试，修复代码或测试用例
2. **Lint错误**：在本地运行Lint工具，修复报告的问题
3. **类型错误**：运行类型检查工具，修复类型不匹配问题
4. **构建失败**：检查构建日志，修复编译问题

9. Fix CI Issues

9. 修复CI问题

For each failed check:
  1. Identify the failure type from check name and logs
  2. Run locally to reproduce:
    • Tests: 
      make test
      or project's test command
    • Lint: 
      make lint
      or project's lint command
    • Types: 
      make typecheck
      or equivalent
  3. Parse error output to identify specific failures
  4. Fix the issues:
    • For test failures: Check if test expectations need updating or if code has a bug
    • For lint errors: Apply automatic fixes where possible, manual fixes otherwise
    • For type errors: Add/fix type annotations or fix type mismatches
  5. Re-run locally to verify fix
  6. Add to task list if not already tracked
Test failure handling:
bash
undefined
针对每个失败的检查：
  1. 从检查名称和日志中识别失败类型
  2. 在本地复现问题：
    • 测试：
      make test
      或项目的测试命令
    • Lint：
      make lint
      或项目的Lint命令
    • 类型检查：
      make typecheck
      或等效命令
  3. 解析错误输出以识别具体失败点
  4. 修复问题：
    • 测试失败：检查是否需要更新测试预期或修复代码bug
    • Lint错误：尽可能应用自动修复，否则手动修复
    • 类型错误：添加/修复类型注解或修复类型不匹配问题
  5. 在本地重新运行以验证修复
  6. 如果未跟踪则添加到任务列表
测试失败处理：
bash
undefined

Run tests and capture output

Run tests and capture output

make test 2>&1 | tee test-output.txt
make test 2>&1 | tee test-output.txt

If using pytest

If using pytest

pytest --tb=short 2>&1 | tee test-output.txt
pytest --tb=short 2>&1 | tee test-output.txt

If using go test

If using go test

go test ./... 2>&1 | tee test-output.txt

Parse output for:
- Failed test names and locations
- Assertion errors with expected vs actual values
- Stack traces pointing to failure source
go test ./... 2>&1 | tee test-output.txt

解析输出内容以获取：
- 失败的测试名称和位置
- 断言错误的预期值与实际值
- 指向失败源的堆栈跟踪

10. Resolve Fixed Threads

10. 标记已修复的线程为已解决

After fixing code-level issues, resolve the corresponding review threads on GitHub:
bash
undefined
修复代码级问题后，在GitHub上标记对应的评审线程为已解决：
bash
undefined

For each fixed code-level thread, resolve it using its thread ID

For each fixed code-level thread, resolve it using its thread ID

gh api graphql -f query=' mutation($threadId: ID!) { resolveReviewThread(input: {threadId: $threadId}) { thread { isResolved } } } ' -f threadId=THREAD_NODE_ID

Only resolve threads whose issues were validated and fixed. Do not resolve threads that were skipped or marked invalid — those need human attention.
gh api graphql -f query=' mutation($threadId: ID!) { resolveReviewThread(input: {threadId: $threadId}) { thread { isResolved } } } ' -f threadId=THREAD_NODE_ID

仅标记已验证并修复的问题对应的线程为已解决。不要标记被跳过或标记为无效的线程——这些需要人工处理。

11. Post Review Report as PR Comment

11. 将评审报告作为PR评论发布

Post the review overview as a comment on the PR. This replaces committing report files.
bash
undefined
将评审概述作为评论发布到PR上。这替代了提交报告文件的方式。
bash
undefined

Post the review overview content as a PR comment

Post the review overview content as a PR comment

gh pr comment $PR_NUM --body "$(cat <<'EOF'
gh pr comment $PR_NUM --body "$(cat <<'EOF'

PR Review Overview - Iteration [N]

PR Review Overview - Iteration [N]

PR: #[number] | Branch: [name] | Date: [YYYY-MM-DD]
PR: #[number] | Branch: [name] | Date: [YYYY-MM-DD]

Valid Issues (fixed)

Valid Issues (fixed)

[List of validated and fixed issues with file:line, reviewer, and brief description]
[List of validated and fixed issues with file:line, reviewer, and brief description]

Invalid/Skipped Issues

Invalid/Skipped Issues

[List of skipped issues with rationale]
[List of skipped issues with rationale]

CI Status

CI Status

[Summary of CI check results and any fixes applied] EOF )"

**Do not commit** `review-overview-*.md` or `review-fixes-*.md` files. They are working files only.
[Summary of CI check results and any fixes applied] EOF )"

**请勿提交** `review-overview-*.md` 或 `review-fixes-*.md` 文件。这些仅为工作文件。

12. Commit, Push, and Verify

12. 提交、推送并验证

After all fixes:
  1. Run full test suite locally
  2. Run linter
  3. Commit only code changes (exclude any local review/task files)
  4. Push to remote
  5. Monitor CI status to confirm checks pass
完成所有修复后：
  1. 在本地运行完整测试套件
  2. 运行Lint工具
  3. 仅提交代码变更（排除任何本地评审/任务文件）
  4. 推送到远程仓库
  5. 监控CI状态以确认检查通过

Key Behaviors

核心行为

  • Auto-fix: Fix all validated issues without pausing for approval
  • Context preservation: Keep diff_hunk context when analyzing
  • Convention adherence: Follow project's existing patterns
  • Deduplication: Consolidate multiple comments on same issue into one task
  • CI verification: Always check CI status after fixing review comments
  • Local reproduction: Run tests/linters locally before pushing fixes
  • Reports as comments: Post review reports as PR comments, never commit them
  • Clean commits: Only commit actual code changes, not working/report files
  • 自动修复：自动修复所有已验证的问题，无需等待批准
  • 上下文保留：分析时保留diff_hunk上下文
  • 遵循规范：遵循项目现有的模式
  • 去重：将同一问题的多条评论合并为一个任务
  • CI验证：修复评审评论后始终检查CI状态
  • 本地复现：推送修复前先在本地运行测试/Lint工具
  • 报告作为评论发布：将评审报告作为PR评论发布，绝不提交到代码库
  • 干净提交：仅提交实际代码变更，不提交工作文件/报告文件