Skill
4
Agent
All Skills
Search
Tools
中文
|
EN
Explore
Loading...
Back to Details
oss-issue-scout
Compare original and translation side by side
🇺🇸
Original
English
🇨🇳
Translation
Chinese
OSS Issue Scout
OSS Issue Scout
Overview
概述
Identify contribution rules and surface a small set of high-value, maintainer-friendly issues for a given GitHub repository.
识别指定GitHub仓库的贡献规则,并筛选出一小部分高价值、符合维护者预期的议题。
Workflow
工作流程
1) Confirm inputs
1) 确认输入信息
Ask for the repo URL or a local repo path and whether GitHub access is available.
Ask for constraints: language/stack familiarity, time budget, desired impact, preferred work type (bug/doc/test/feature), and skill level.
Ask if the user can run tests and open PRs.
询问用户仓库URL或本地仓库路径,以及是否具备GitHub访问权限。
询问约束条件:熟悉的技术栈/编程语言、时间预算、期望影响、偏好的工作类型(Bug修复/文档/测试/功能开发),以及技能水平。
询问用户是否能够运行测试并提交PR。
2) Read contribution rules and signals
2) 读取贡献规则与相关信号
Check repo docs locally or via GitHub: CONTRIBUTING.md, README, CODE_OF_CONDUCT, SECURITY.md, LICENSE, GOVERNANCE, MAINTAINERS, issue/PR templates, .github/ folder, CI configs.
Note CLA/DCO requirements, branch naming, commit rules, test expectations, and required checks.
If any open-source or GitHub practice is unclear, use Context7 to refresh before advising.
本地或通过GitHub查看仓库文档:CONTRIBUTING.md、README、CODE_OF_CONDUCT、SECURITY.md、LICENSE、GOVERNANCE、MAINTAINERS、议题/PR模板、.github/文件夹、CI配置文件。
记录CLA/DCO要求、分支命名规范、提交规则、测试要求以及必要的检查项。
如果对任何开源或GitHub实践有疑问,先通过Context7查阅相关信息再提供建议。
3) Pull issues via GitHub MCP
3) 通过GitHub MCP拉取议题
Discover available GitHub MCP resources/templates and use them to list open issues with labels, assignees, milestones, and last activity.
Prefer issues labeled good first issue/help wanted/bug/docs/tests and those with clear acceptance criteria.
Filter out: stale/needs info, blocked, huge refactors, security-sensitive items, no reproduction steps, or already assigned issues.
Cross-check for linked PRs, duplicates, or active discussions that indicate ongoing work.
查找可用的GitHub MCP资源/模板,并用其列出带有标签、经办人、里程碑和最后活动时间的开放议题。
优先选择带有good first issue/help wanted/bug/docs/tests标签,且验收标准明确的议题。
排除以下议题:长期未更新/需要补充信息、被阻塞、大型重构、安全敏感、无复现步骤,或是已被指派的议题。
交叉检查是否有关联的PR、重复议题,或是表明正在进行中的活跃讨论。
4) Score and shortlist
4) 评分与筛选
Use
references/issue-scoring.md
to score value, acceptance likelihood, clarity, and effort fit.
Select 3-7 issues with the best balance of impact and low acceptance risk; keep scope aligned to user constraints.
使用
references/issue-scoring.md
对议题的价值、被接受可能性、清晰度以及工作量匹配度进行评分。
选择3-7个在影响力与低验收风险之间达到最佳平衡的议题;确保范围符合用户的约束条件。
5) Report back
5) 反馈结果
Provide a concise list of recommended issues with links, labels, last activity, estimated effort, and why they are good picks.
Summarize contribution rules and any blockers.
Ask targeted questions if key info is missing or if all candidate issues are low quality.
提供简洁的推荐议题列表,包含链接、标签、最后活动时间、预估工作量,以及推荐理由。
总结贡献规则与任何限制条件。
如果关键信息缺失,或是所有候选议题质量较低,提出针对性问题。
Output format
输出格式
Recommended issues
Contribution rules and constraints
Risks/unknowns
Questions for the user
Suggested next actions
推荐议题
贡献规则与约束条件
风险/未知项
向用户提出的问题
建议的后续行动
References
参考资料
Read
references/issue-scoring.md
when scoring or filtering issues.
评分或筛选议题时,请查阅
references/issue-scoring.md
。