ONES Doc MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| ONES_COOKIE | No | 浏览器会话里的 Cookie，会话复用模式可选但建议提供 | |
| ONES_ORIGIN | No | 请求使用的 Origin，可选，默认回落到 ONES_BASE_URL | |
| ONES_REFERER | No | 请求使用的 Referer，可选，默认回落到 ${ONES_BASE_URL}/project/ | |
| ONES_BASE_URL | No | ONES 内网根地址 | |
| ONES_PASSWORD | No | 登录密码，账号密码模式必填 | |
| ONES_USERNAME | No | 登录账号，账号密码模式必填 | |
| ONES_AUTH_TOKEN | No | 浏览器会话里的 Bearer token，会话复用模式可选但建议提供 | |
| ONES_TIMEOUT_MS | No | 请求超时，可选，默认 15000 | |
| ONES_USER_AGENT | No | 请求使用的 User-Agent，可选；账号密码模式下未配置时默认使用企微风格 UA | |
| ONES_OCR_API_KEY | No | OCR 服务鉴权 token，可选 | |
| ONES_OCR_ENDPOINT | No | OCR 服务地址，可选 | |
| ONES_OCR_PROVIDER | No | OCR 提供方，可选，当前支持 http | |
| ONES_OCR_TIMEOUT_MS | No | OCR 请求超时，可选，默认 15000 | |
| ONES_MAX_CONTENT_CHARS | No | 正文最大长度，可选，默认 20000 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| search_docsC | Search ONES docs by keyword |
| get_docA | Get an ONES document by URL or #requirement. With include_resources=true, refresh image URLs and return stable resource_id metadata; then call download_ones_resource_by_id for MCP-native image content. |
| get_doc_outlineA | Return section outline and size metadata for a ONES doc ref. |
| get_doc_sectionB | Return one outline section from a ONES doc ref, optionally including descendants. |
| get_doc_chunksB | Return one cursor-based chunk from a ONES doc ref using a character budget. |
| get_doc_contextB | Automatically select sections or chunks for a ONES doc ref based on a question. |
| get_requirement_detail_by_refA | Get requirement body, fields, and related task facts by requirement number, task id, or task URL. |
| get_execution_tasks_by_refA | Get execution-task candidates by requirement number, task id, or task URL. |
| extract_requirement_materials_by_refA | Extract wiki pages, external links, rich resources, and completeness hints by requirement number, task id, or task URL. |
| list_requirement_bugs_by_refA | List bugs under a requirement by requirement number, task id, or task URL when explicitly requested. |
| get_task_messages_by_refB | Get task comments/messages by task number, task id, or task URL. |
| get_related_wiki_pages_by_refB | Discover wiki page references by requirement number, task id, or task URL. |
| get_task_rich_resources_by_refB | Extract rich-text image resources by task number, task id, or task URL. |
| get_bug_detail_by_refA | Get bug body, fields, and related task facts by bug number, task id, or task URL. |
| get_bug_parent_requirement_by_refB | Resolve the parent requirement by bug number, task id, or task URL. |
| resolve_requirementA | Resolve an ONES requirement by number, task id, or task URL. Prefer *_by_ref tools for normal workflow reads. |
| get_requirement_detailA | Get requirement body, fields, and related task facts by task id. Prefer get_requirement_detail_by_ref when starting from a requirement number or URL. |
| get_execution_tasksA | Get execution-task candidates related to a requirement by task id. Prefer get_execution_tasks_by_ref when starting from a requirement number or URL. |
| resolve_bugA | Resolve an ONES bug by number, task id, or task URL. Prefer *_by_ref tools for normal workflow reads. |
| get_bug_detailA | Get bug body, fields, and related task facts by task id. Prefer get_bug_detail_by_ref when starting from a bug number or URL. |
| get_bug_parent_requirementA | Resolve the parent requirement for a selected bug by task id. Prefer get_bug_parent_requirement_by_ref when starting from a bug number or URL. |
| list_requirement_bugsA | List bugs under a requirement by task id when explicitly requested. Prefer list_requirement_bugs_by_ref when starting from a requirement number or URL. |
| get_task_messagesA | Get task comments/messages by task id for requirement workflow context. Prefer get_task_messages_by_ref when starting from a number or URL. |
| extract_requirement_materialsA | Extract wiki pages, external links, rich resources, and completeness hints from a requirement task id. Prefer extract_requirement_materials_by_ref when starting from a requirement number or URL. |
| get_related_wiki_pagesA | Discover wiki page references attached to or linked from a requirement task id. Prefer get_related_wiki_pages_by_ref when starting from a requirement number or URL. |
| get_task_rich_resourcesA | Extract rich-text image resources from a requirement, task, or bug body by task id. Prefer get_task_rich_resources_by_ref when starting from a number or URL. |
| download_ones_resourceA | Download an ONES resource by URL. Signed image URLs can expire; prefer get_doc(include_resources=true) followed by download_ones_resource_by_id. |
| download_ones_resource_by_idA | Resolve a fresh signed URL for a stable ONES resource_id and return the image as MCP ImageContent. Recommended flow: get_doc(include_resources=true), read resources, then call this tool. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 28 tools
The server is dominated by near-duplicate tool pairs (e.g. get_requirement_detail vs get_requirement_detail_by_ref, get_bug_detail vs get_bug_detail_by_ref, get_task_messages vs get_task_messages_by_ref) that differ only in whether the input is a task id or a number/URL. The descriptions explicitly state 'prefer *_by_ref for normal workflow reads,' which helps an agent choose, but the systematic duplication still creates real misselection risk. The doc tools (get_doc, get_doc_section, get_doc_chunks, get_doc_outline, get_doc_context) are mostly distinct by granularity, though get_doc_context overlaps with section/chunk retrieval.
All tools use snake_case verb_noun phrasing (get_, resolve_, list_, extract_, search_, download_) with a consistent '_by_ref' suffix convention for the reference-based variants. Minor stylistic drift exists between get_ vs resolve_ for related operations, but the pattern is largely predictable.
At 28 tools the set is heavy for the stated scope, and much of the bloat comes from duplicating roughly ten read operations into non-ref/ref pairs. Collapsing these pairs would bring the surface to a much more reasonable ~15 tools, so the current count is inflated rather than earned.
Coverage spans docs (retrieval, outline, chunks, search, resource download), requirements, bugs, tasks, comments, wiki pages, and rich resources, with by_ref variants for most read paths. It is a read-oriented surface with no create/update operations, but for a documentation/inspection server the lifecycle coverage is solid with only minor gaps (e.g. no generic task-detail fetch, search limited to docs).