# Knowledge Base The knowledge base is a set of markdown files the daemon indexes at startup and makes available to agents and users via search, MCP tools, and [MCP resources](https://modelcontextprotocol.io/specification/2025-11-25/server/resources). ## Source directories Three directories, each with a distinct role: | Directory | Purpose | |-----------|---------| | `wiki/` | How the org works, architecture, decisions | | `blueprints/` | What correct looks like | | `recipes/` | Step-by-step how to do things | Drop a markdown file in any of these directories and restart the daemon. ## Frontmatter All fields are optional; a file with no frontmatter indexes fine. ```yaml --- title: Git Workflow # display title; defaults to the first H1, then the filename slug editable: true # allow propose_edit to target this doc (default false) category: conventions # tree position under the doc's URI scheme — single level tags: [git, process] # cross-cutting topic labels, lowercase-kebab --- ``` `category` groups docs into a branch in `bento kb`; uncategorized docs sit at the scheme root. `tags` render as a suffix in the listing, filter via `bento kb --tag `, and appear in the doc's MCP resource description. ## Method & delivery How knowledge reaches an agent run is configured per pipeline with the [`knowledge:` block](/pipelines/config#knowledge): two knobs — `method` (how the daemon retrieves) and `delivery` (how results reach the agent), composable with a `search` escape hatch. See [Method & delivery](/knowledge-base/modes) for the deep dive. ## Querying the knowledge base Bento offers three ways to query the knowledge base. 1. Via MCP on your harness (claude, codex, etc) ```bash (claude) "how do we handle git branches" => Returns LLM response ``` 2. Via the CLI using the `bento ask` command ```bash bento ask "how do we handle git branches" bento ask "how do we handle git branches" --fast => Returns LLM response ``` 3. Directly as MCP resources\*\* :::warning \*\*Caution: MCP resources are invisibly injected into the context when connected as `wiki://`, `blueprint://`, and `recipe://` URIs. Something to be aware of when querying of from a context pollution perspective, but also a powerful way to structure and link knowledge. ::: ```bash "Do we have any git resources in the wiki?" => Returns LLM list of wiki resources, e.g.: - wiki://git-workflow - wiki://git-commit-message-standards ``` ## Why the separation? Create a fast feedback loop to ensure knowledge can be retrieved consistently for both Humans and Agents. If you are unable to retrieve a known document to statement via any of these methods, then it's also unlikely that Agent will be unable to surface and use that knowledge when it counts (e.g. mid-run in a pipeline). ## Notes * **Docs are indexed at startup, not on save.** Adding or editing a file requires `bento daemon restart` to take effect. * **Prefer prose over tables for searchable content.** Tables and bullet lists embed poorly as standalone chunks; a prose sentence covering the same convention surfaces more reliably for natural-language queries. * See [MCP](/knowledge-base/mcp) for detailed information.