Documentation MCP response
MCP tool response returned by the Documentation MCP tool.
- Returned by: GovTribe Documentation MCP tool
- Response type: Structured documentation search and retrieval results
Response Shape
The response combines a text markdown answer with structured content. Agents should use the structured fields to discover article selectors and use content_markdown when they need the retrieved documentation text.
| Field | Type | Notes |
|---|---|---|
query | string or null | The search query used for discovery. Exact article, URL, or part retrieval returns null. |
collections | string array | Collection filters applied to the request. Empty means the search was not narrowed by collection. |
available_collections | object | Collection slugs and labels that agents can use in later collections filters. These values match the collections enum exposed by the tool schema. |
result_count | integer | Number of returned documentation segments. |
truncated | boolean | Whether the response hit the request max_tokens budget before all matching content could be included. |
content_markdown | string | Concatenated markdown content returned within the token budget. |
results | object array | Matching documentation segments with article selectors, metadata, excerpts, scores, and optional per-segment markdown. |
Each results item includes article_name, article_url, segment_url, collection, collection_label, breadcrumbs, part_id, segment_type, heading, excerpt, score, estimated_tokens, truncated, and content_markdown.
article_name, article_url, and part_id are retrievable selectors. Agents can pass them back to Documentation as article_names, article_urls, or part_ids when they need exact content.
Follow-up Behavior
Use include_content: false for discovery-only searches when the agent only needs article names, URLs, collections, and part IDs. Use max_tokens to control how much documentation markdown comes back.
When truncated is true, call Documentation again with a larger max_tokens value or with a narrower selector such as one returned part_id. When an agent needs a whole article, retrieve by article_names or article_urls; when it needs only one section, retrieve by part_ids.
Example
{
"query": "saved search",
"collections": [],
"available_collections": {
"data-model": "Data model",
"govtribe-for-agents": "GovTribe for Agents",
"govtribe-user-guide": "GovTribe User Guide"
},
"result_count": 2,
"truncated": false,
"content_markdown": "### Saved searches\n\nSource: /docs/govtribe-user-guide/capture/saved-searches\nPart: govtribe-user-guide/capture/saved-searches::saved-searches-01\n\nA short retrieved documentation excerpt.",
"results": [
{
"article_name": "Saved searches",
"article_url": "/docs/govtribe-user-guide/capture/saved-searches",
"segment_url": "/docs/govtribe-user-guide/capture/saved-searches",
"collection": "govtribe-user-guide",
"collection_label": "GovTribe User Guide",
"breadcrumbs": [
"GovTribe documentation",
"GovTribe User Guide",
"Capture"
],
"part_id": "govtribe-user-guide/capture/saved-searches::saved-searches-01",
"segment_type": "page",
"heading": "Saved searches",
"excerpt": "A short matched excerpt from the Saved searches article.",
"score": 45.0556,
"estimated_tokens": 52,
"truncated": false,
"content_markdown": "A short retrieved documentation excerpt."
},
{
"article_name": "Saved searches",
"article_url": "/docs/govtribe-user-guide/capture/saved-searches",
"segment_url": "/docs/govtribe-user-guide/capture/saved-searches#row-actions",
"collection": "govtribe-user-guide",
"collection_label": "GovTribe User Guide",
"breadcrumbs": [
"GovTribe documentation",
"GovTribe User Guide",
"Capture"
],
"part_id": "govtribe-user-guide/capture/saved-searches::row-actions-07",
"segment_type": "heading",
"heading": "Row actions",
"excerpt": "A short matched excerpt from the Row actions section.",
"score": 42.1327,
"estimated_tokens": 208,
"truncated": false,
"content_markdown": "A short retrieved section excerpt."
}
]
}