菜单
Use when synthesizing information from multiple research sources (papers, repos, blogs) into a coherent knowledge base. Triggers when multiple references have been gathered and need to be integrated, deduplicated, or cross-referenced.
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
Use when searching for academic papers, research references, or scientific literature on arxiv.org. Triggers when user asks to find papers, citations, related work, or state-of-the-art research on a topic.
Interact with Confluence Cloud via REST API — create pages, update content, search with CQL, read documentation, and manage spaces. Use this skill whenever the user mentions Confluence pages, wiki, documentation, knowledge base, spaces, or anything related to Confluence content management, even if they don't explicitly say 'Confluence'.
Use when searching for codebase architecture, structure, or documentation of open-source GitHub repositories via DeepWiki. Triggers when user asks to understand a repo's architecture, internals, codebase structure, or how a specific subsystem works in an open-source project.
Use when searching for open-source implementations, code repositories, or reference codebases on GitHub. Triggers when user asks to find code, implementations, libraries, or projects related to a technical topic.
Interact with Jira Cloud via REST API — create issues, search with JQL, view details, update fields, transition workflow states, add comments, list projects, and browse boards/sprints. Use this skill whenever the user mentions Jira tickets, issues, sprints, boards, JQL, or anything related to Jira project management, even if they don't explicitly say 'Jira'.
Automatically fill PR template from plan.md, commit, push, and open a GitHub pull request for the current branch. Use when the user says "open pr", "create pr", "submit pr", or any variation of opening a pull request.
| name | information-combiner |
| description | Use when synthesizing information from multiple research sources (papers, repos, blogs) into a coherent knowledge base. Triggers when multiple references have been gathered and need to be integrated, deduplicated, or cross-referenced. |
Synthesize findings from multiple sources into a unified, coherent knowledge structure.
List every source collected with type tag:
[PAPER] Title — arxiv:2301.00001
[CODE] owner/repo — github.com/...
[BLOG] Post Title — url
For each source, extract along these axes:
| Dimension | Questions to Answer |
|---|---|
| Core claim | What is the main contribution/finding? |
| Method | How does it work? Key techniques? |
| Results | Benchmarks, metrics, comparisons |
| Limitations | What doesn't it solve? |
| Dependencies | What prior work does it build on? |
| Date | When was this published/updated? |
Build a comparison matrix for the key sources:
| Aspect | Source A | Source B | Source C |
|--------------|----------|----------|----------|
| Method | ... | ... | ... |
| Dataset | ... | ... | ... |
| Performance | ... | ... | ... |
| Code avail. | ✓ | ✗ | ✓ |
Mark relationships between sources:
After mapping sources, identify:
## Combined Research Summary: [Topic]
### Key Themes
1. [Theme 1]: [Sources that cover this] — [consensus finding]
2. [Theme 2]: [Sources] — [finding]
### Source Relationships
- [Paper A] → implemented by → [Repo B]
- [Paper A] → explained by → [Blog C]
- [Paper D] → improves upon → [Paper A]
### Comparison Matrix
| ... | ... |
### Consensus Findings
- [Finding 1]: Agreed upon by [Source A, B, C]
- [Finding 2]: ...
### Contradictions / Open Questions
- [Contradiction]: [Source A] claims X, [Source B] claims Y
- [Open question]: No source covers Z
### Recommended Reading Order
1. [Blog C] — best intuitive introduction
2. [Paper A] — foundational method
3. [Repo B] — practical implementation