ワンクリックで
doc-read
Search and retrieve content from the doctree-mcp knowledge base. Use when you need to find information in the indexed markdown docs.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Search and retrieve content from the doctree-mcp knowledge base. Use when you need to find information in the indexed markdown docs.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Audit the doctree-mcp wiki for orphaned pages, stubs, missing frontmatter, and broken links.
Create a new wiki entry in the doctree-mcp knowledge base. Checks for duplicates, generates a scaffold, and writes with validation. Requires WIKI_WRITE=1.
SOC 職業分類に基づく
| name | doc-read |
| description | Search and retrieve content from the doctree-mcp knowledge base. Use when you need to find information in the indexed markdown docs. |
| argument-hint | ["search query or topic"] |
Find and retrieve content from the markdown knowledge base using the doctree-mcp tools.
Follow these steps in order. Do NOT skip the tree step — it lets you target exactly what to read instead of dumping entire documents.
If you have a specific query, search for it:
search_documents(query: "$ARGUMENTS")
If you're exploring broadly or want to see what's available:
list_documents(query: "$ARGUMENTS")
Note the doc_id and node_id values from the results.
Pick the most relevant document from step 1 and get its structure:
get_tree(doc_id: "<doc_id from step 1>")
Read the outline carefully. Identify which sections are relevant to the query based on their titles, word counts, and summaries. Do NOT retrieve everything — pick only what matters.
For a single section or a few sections:
get_node_content(doc_id: "<doc_id>", node_ids: ["<node_id>", ...])
For a section and all its subsections (more efficient than fetching each child):
navigate_tree(doc_id: "<doc_id>", node_id: "<parent_node_id>")
If the retrieved content mentions or links to other documents, repeat steps 2-3 for those documents if the user needs that information.
filters: { "type": "runbook", "tags": ["auth"] }navigate_tree over multiple get_node_content calls when you need a whole section with children