深度话题调研工作流
AdvisoryAudited by Static analysis on May 4, 2026.
Overview
No suspicious patterns detected.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent may visit external websites and summarize their content while preparing the report.
The workflow expects browser or web extraction tools to fetch external pages. This is central to the stated research purpose and bounded by the instruction to deep-read only 1-3 relevant pages.
用 `browser_navigate` 或 `web_extract` 获取页面全文
Review cited sources and avoid asking the skill to research or transmit sensitive private information unless necessary.
The research topic and relevant context may be passed to delegated agents during parallel search.
The skill instructs the agent to delegate parts of the research to subagents. This is disclosed and purpose-aligned, but the artifact does not specify subagent identity or data-boundary details.
用 `delegate_task` 并行搜索，每个子智能体负责一个维度。
Do not include secrets, confidential business information, or unnecessary personal data in the research prompt.
Generated research notes may remain in the user's knowledge base and be reused later, including any inaccuracies if sources were not verified.
The workflow can persist the generated research report into an Obsidian note when requested by the user. This is user-directed and scoped by a user-specified path.
方式 B：写入Obsidian（如用户要求） 文件路径由用户在 Phase 0 指定。
Only choose the Obsidian option when you want a persistent note, and verify important claims before relying on or reusing the saved report.