BW-agent--knowledge

Use when a user provides a source document (ChatGPT export, article, PRD, transcript) and asks to extract knowledge points and integrate them into the Obsidian knowledge vault. Trigger phrases: 这份文档涉及的知识点, 写入工程知识库, 集成到vault, 知识提取.

当用户提供源文档（如 ChatGPT 导出内容、文章、产品需求文档、会议转录等），并请求从中提取关键知识点并将其整合至 Obsidian 知识库时（例如将这份文档涉及的知识点, 写入工程知识库, 集成到vault, 知识提取），请使用此skill。

Map source code modules to Obsidian knowledge-tree nodes and generate source-evidence mapping.

Maintain Obsidian `_index.md` files by linking every Markdown note in each branch folder.

Generate interview question indexes from the current Vault structure. Do not generate answers unless explicitly requested.

Build meaningful Obsidian double links across the Vault.

Normalize Obsidian Markdown notes by adding missing frontmatter and required headings without overwriting existing content.

Generate source-based project evidence cards for interview expression.

Map AI / Agent knowledge points to the user's projects using only known project facts.

Audit a source code repository and summarize stack, structure, core modules, entry points, execution chains, and interview evidence. Read-only.

Generate source-code-based interview deep-dive questions.

Identify source-code gaps and interview risks by comparing claimed capabilities with actual implementation evidence.

Generate resume-safe bullets from source code evidence without inventing metrics.

Audit Obsidian Vault structure, metadata, templates, links, and coverage. Read-only by default.

Map source code modules to Obsidian knowledge-tree nodes and generate source-evidence mapping.

Maintain Obsidian `_index.md` files by linking every Markdown note in each branch folder.

Generate interview question indexes from the current Vault structure.

Build and repair meaningful Obsidian double links across the Vault.

Normalize Obsidian Markdown note templates by adding missing frontmatter and required headings without overwriting existing content.

Generate source-based project evidence cards for interview expression.

Map AI / Agent knowledge points to the user projects using only known facts.

Audit a source code repository and summarize stack, structure, core modules, entry points, execution chains, and interview evidence. Read-only.

Generate interview deep-dive questions from source code modules.

Identify source-code gaps and interview risks by comparing claimed capabilities with actual implementation evidence.

Generate resume bullets from source code evidence without inventing metrics.

Audit an Obsidian Vault for structure, metadata, templates, links, and coverage. Read-only by default.