Read and export password-protected ShowDoc projects as Markdown via CLI, HTTP, or MCP.
ShowDoc2MD demonstrates solid tool design with clear naming (verb_noun pattern), comprehensive parameter descriptions, and well-documented schemas. All 5 tools follow consistent patterns and include type information. However, output schemas lack formal documentation in code, and error handling could be more granular. Tool descriptions are adequate (100-200 chars) but could better explain when to use each tool vs. alternatives. Parameter descriptions are thorough and include constraints (e.g., verify_ssl default=true). The server correctly avoids exposing secrets as parameters (uses SHOWDOC_PASSWORD env var). Main gaps: no tool annotations (readOnlyHint/destructiveHint), no pagination support for list_pages, and error messages are in Chinese (limits LLM understanding in English contexts).
Export a ShowDoc project to Markdown files on the MCP server machine. Returns paths on the MCP server machine. The result includes complete=false if any page or requested asset failed.
List every page in a ShowDoc project without downloading page bodies.
Check whether a ShowDoc URL can be read with an already-known document password.
Read the whole ShowDoc project and return one combined Markdown string. Prefer showdoc_list_pages + showdoc_read_page for large projects when only a few pages are needed. This tool does not persist files and does not download image bytes.
Read one ShowDoc page and return Markdown suitable for an AI context. If page_id is omitted, the page ID embedded in the supplied URL is used.
Output schemas not formally documented in code. Return types are inferred from code but not declared in tool registration. LLMs cannot see what fields to expect without reading implementation.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). showdoc_export is destructive (writes files) but unmarked. LLMs cannot distinguish safe reads from state-modifying operations.
Error messages in Chinese (e.g., '缺少 ShowDoc 文档密码') limit LLM understanding in English-only contexts. Errors should be in English or support i18n.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 66 | 2026-07-28+ | v2 |
showdoc_list_pages returns all pages without pagination. Large projects could return hundreds of items, bloating context. Should support limit/offset and return total_count.
showdoc_read_full lacks dry-run or confirmation step. Agents could accidentally export large projects to disk without explicit user approval.