一键导入
sp-address-pr-comments
Address and resolve existing review comments on a GitHub pull request.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Address and resolve existing review comments on a GitHub pull request.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | sp-address-pr-comments |
| description | Address and resolve existing review comments on a GitHub pull request. |
| mode | agent |
| tools | ["codebase","githubRepo","terminalLastCommand"] |
| arguments | [{"name":"pr_number","description":"The number of the GitHub pull request to address comments on.","required":true}] |
Address and resolve existing comments/feedback raised on this PR:
Retrieve the review comments and threads for the pull request to identify outstanding feedback:
pull_request_read tool with method = "get_review_comments", specifying the repository owner, repo, and the pull request pullNumber.ID (required to post replies later).is_resolved. Only address unresolved threads.path, and code line number.Before touching any file, check its current line count:
wc -l <path/to/file.kt>
Ensure you adhere to the project quality and style standards, including detekt rules, ktlint formatting, test coverage requirements, and the 600-line file-size limit.
Do not consider the comments addressed until ALL of the following pass locally:
make check # Konsist architecture rules + all unit tests + debug compilation — must show 0 failures
Once all changes are implemented and verified, you MUST reply to every comment/thread on GitHub using the add_reply_to_pull_request_comment tool from the GitHub MCP server:
Fetch, analyze, and resolve SonarQube quality gate failures, code smells, bugs, and coverage gaps.
Turn an idea, problem description, or bug report into a well-formed GitHub issue and create it on GitHub once approved.
Study the feasibility and viability of a proposed feature or architectural change.
Request follow-up work on a specific issue (e.g. bug fixes, additional requirements, visual regression, or debugging) in a new or existing conversation without prior context.
Use for any question about a codebase, its architecture, file relationships, or project content — especially when graphify-out/ exists, where the question should be treated as a graphify query first. Turns any input (code, docs, papers, images, videos) into a persistent knowledge graph with god nodes, community detection, and query/path/explain tools.
Plan a GitHub issue by investigating context and creating a detailed implementation plan. Present it to the user, and post it to the issue as a comment upon approval. No code changes.