Tool reference · Save & recall
memory_context_build
Assembles a token-bounded context bundle for a task: related knowledge, recent episodes, applicable skills, rules, and known blind spots.
local stdio server read-only
When to use
- You are starting a larger task and want one bundle of relevant background instead of several searches.
- You need the context to fit a fixed token budget.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
queryrequired | string | — | What you need context for |
project | string | — | — |
max_tokens | integer | 4000 | — |
Example
Arguments
{
"query": "refactor billing job scheduling",
"project": "my-api",
"max_tokens": 3000
}Result shape
{
"knowledge": [
{
"id": 1842,
"type": "decision",
"content": "Use PostgreSQL advisory locks for the nightly billing job…",
"project": "my-api",
"activation_score": 0.72
}
],
"episodes": [],
"skills": [],
"rules": [],
"competency": null,
"blind_spots": [],
"total_tokens": 312
}Episodes and blind spots are only filled when project is given.
Values are illustrative; the keys follow the server's handler. MCP clients receive the result as JSON text content.
Server description
The description the server sends to your agent in tools/list, captured from the v14.7.0 source:
Build optimal context for a query. Combines: spreading activation + knowledge graph + episodes + skills + self-model. The 'brain thinking' tool.