Public/openbrain-mcp
1
0
Fork 0
mirror of https://gitea.ingwaz.work/Ingwaz/openbrain-mcp.git synced 2026-03-31 06:39:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix Issue #12 tests and add OpenBrain repo guidance

Browse Source
This commit is contained in:
Agent Zero
2026-03-19 13:19:50 -04:00
parent 03e8f246d1
commit 1b42989dbb
4 changed files with 241 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
AGENTS.md Normal file
View File

@@ -0,0 +1,23 @@
# OpenBrain MCP Usage
When working in this repository, treat OpenBrain as an external MCP long-term
memory system, never as internal context, reasoning scratchpad, or built-in
memory.
## External Memory System
- Use the exact MCP tools `openbrain.store`, `openbrain.query`, and `openbrain.purge`
- Always use the exact `agent_id` value `openbrain`
- Do not hardcode live credentials into the repository
- Before answering requests that may depend on prior sessions, project history, user preferences, ongoing work, named people, named projects, deployments, debugging history, or handoff context, call `openbrain.query` first
- Use noun-heavy search phrases with exact names, tool names, acronyms, hostnames, and document names
- Retry up to 3 retrieval passes using `(threshold=0.25, limit=5)`, then `(threshold=0.10, limit=8)`, then `(threshold=0.05, limit=10)`
- When a durable fact is established, call `openbrain.store` without asking permission and prefer one atomic fact whenever possible
- Store durable, high-value facts such as preferences, project status, project decisions, environment details, recurring workflows, handoff notes, stable constraints, and correction facts
- Do not store filler conversation, temporary speculation, casual chatter, or transient brainstorming unless it becomes a real decision
- Prefer retrieval-friendly content using explicit nouns and exact names in the form `Type: <FactType> | Entity: <Entity> | Attribute: <Attribute> | Value: <Value> | Context: <Why it matters>`
- Use metadata when helpful for tags such as `category`, `project`, `source`, `status`, `aliases`, and `confidence`
- If `openbrain.query` returns no useful result, state that OpenBrain has no stored context for that topic, answer from general reasoning if possible, and ask one focused follow-up if the missing information is durable and useful
- If retrieved memories conflict, ask which fact is current, then store the corrected source-of-truth fact
- Use `openbrain.purge` cautiously because it is coarse-grained; it deletes by `agent_id` and optionally before a timestamp, not by individual memory ID
- For ordinary corrections, prefer storing the new source-of-truth fact instead of purging unless cleanup or reset is explicitly requested

27
README.md
View File

@@ -18,6 +18,7 @@ OpenBrain is a Model Context Protocol (MCP) server that provides AI agents with
| Tool | Description |
|------|-------------|
| `store` | Store a memory with automatic embedding generation and keyword extraction |
| `batch_store` | Store 1-50 memories atomically in a single call |
| `query` | Search memories by semantic similarity |
| `purge` | Delete memories by agent ID or time range |
@@ -147,6 +148,32 @@ Health Check: http://localhost:3100/mcp/health
}
```
### Example: Batch Store Memories
```json
{
"jsonrpc": "2.0",
"id": 3,
"method": "tools/call",
"params": {
"name": "batch_store",
"arguments": {
"agent_id": "assistant-1",
"entries": [
{
"content": "The user prefers dark mode",
"metadata": {"category": "preference"}
},
{
"content": "The user uses vim keybindings",
"metadata": {"category": "preference"}
}
]
}
}
}
```
## Architecture
```

74
src/tools/mod.rs
View File

@@ -15,31 +15,50 @@ pub fn get_tool_definitions() -> Vec<Value> {
vec![
json!({
"name": "store",
"description": "Store a memory with automatic embedding generation",
"description": "Store a memory with automatic embedding generation and keyword extraction. The memory will be associated with the agent_id for isolated retrieval.",
"inputSchema": {
"type": "object",
"properties": {
"content": {"type": "string"},
"agent_id": {"type": "string"},
"metadata": {"type": "object"}
"content": {
"type": "string",
"description": "The text content to store as a memory"
},
"agent_id": {
"type": "string",
"description": "Unique identifier for the agent storing the memory (default: 'default')"
},
"metadata": {
"type": "object",
"description": "Optional metadata to attach to the memory"
}
},
"required": ["content"]
}
}),
json!({
"name": "batch_store",
"description": "Store multiple memories in a single call (1-50 entries)",
"description": "Store multiple memories with automatic embedding generation and keyword extraction. Accepts 1-50 entries and stores them atomically in a single transaction.",
"inputSchema": {
"type": "object",
"properties": {
"agent_id": {"type": "string"},
"agent_id": {
"type": "string",
"description": "Unique identifier for the agent storing the memories (default: 'default')"
},
"entries": {
"type": "array",
"description": "Array of 1-50 memory entries to store atomically",
"items": {
"type": "object",
"properties": {
"content": {"type": "string"},
"metadata": {"type": "object"}
"content": {
"type": "string",
"description": "The text content to store as a memory"
},
"metadata": {
"type": "object",
"description": "Optional metadata to attach to the memory"
}
},
"required": ["content"]
}
@@ -50,27 +69,48 @@ pub fn get_tool_definitions() -> Vec<Value> {
}),
json!({
"name": "query",
"description": "Query memories by semantic similarity",
"description": "Query stored memories using semantic similarity search. Returns the most relevant memories based on the query text.",
"inputSchema": {
"type": "object",
"properties": {
"query": {"type": "string"},
"agent_id": {"type": "string"},
"limit": {"type": "integer"},
"threshold": {"type": "number"}
"query": {
"type": "string",
"description": "The search query text"
},
"agent_id": {
"type": "string",
"description": "Agent ID to search within (default: 'default')"
},
"limit": {
"type": "integer",
"description": "Maximum number of results to return (default: 10)"
},
"threshold": {
"type": "number",
"description": "Minimum similarity threshold 0.0-1.0 (default: 0.5)"
}
},
"required": ["query"]
}
}),
json!({
"name": "purge",
"description": "Delete memories by agent_id",
"description": "Delete memories for an agent. Can delete all memories or those before a specific timestamp.",
"inputSchema": {
"type": "object",
"properties": {
"agent_id": {"type": "string"},
"before": {"type": "string"},
"confirm": {"type": "boolean"}
"agent_id": {
"type": "string",
"description": "Agent ID whose memories to delete (required)"
},
"before": {
"type": "string",
"description": "Optional ISO8601 timestamp - delete memories created before this time"
},
"confirm": {
"type": "boolean",
"description": "Must be true to confirm deletion"
}
},
"required": ["agent_id", "confirm"]
}

164
tests/e2e_mcp.rs
View File

@@ -879,60 +879,155 @@ async fn e2e_auth_enabled_accepts_test_key() {
#[tokio::test]
async fn e2e_batch_store_basic() -> anyhow::Result<()> {
let agent = format!("batch_{}", uuid::Uuid::new_v4());
let _ = db.purge_memories(&agent, None).await;
let base = base_url();
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(20))
.build()
.expect("reqwest client");
let resp = client.call_tool("batch_store", serde_json::json!({
"agent_id": agent.clone(),
ensure_schema().await;
wait_until_ready(&client, &base).await;
let agent = format!("batch_{}", uuid::Uuid::new_v4());
let _ = call_tool(&client, &base, "purge", json!({ "agent_id": agent, "confirm": true })).await;
let result = call_tool(&client, &base, "batch_store", serde_json::json!({
"agent_id": agent,
"entries": [
{"content": "Fact alpha for batch test"},
{"content": "Fact beta for batch test"},
{"content": "Fact gamma for batch test"}
]
})).await?;
})).await;
let result: Value = serde_json::from_str(&resp.content[0].text)?;
assert!(result["success"].as_bool().unwrap_or(false));
assert_eq!(result["count"].as_i64().unwrap_or(0), 3);
db.purge_memories(&agent, None).await?;
Ok(())
}
#[tokio::test]
async fn e2e_batch_store_empty_rejected() -> anyhow::Result<()> {
let resp = client.call_tool("batch_store", serde_json::json!({
"entries": []
})).await;
assert!(resp.is_err() || resp.as_ref().unwrap().is_error());
let base = base_url();
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(20))
.build()
.expect("reqwest client");
wait_until_ready(&client, &base).await;
let response = call_jsonrpc(
&client,
&base,
json!({
"jsonrpc": "2.0",
"id": "batch-empty-1",
"method": "tools/call",
"params": {
"name": "batch_store",
"arguments": {
"entries": []
}
}
}),
)
.await;
assert!(response.get("error").is_some(), "empty batch_store should return an error");
Ok(())
}
#[tokio::test]
async fn e2e_batch_store_exceeds_max() -> anyhow::Result<()> {
let base = base_url();
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(20))
.build()
.expect("reqwest client");
wait_until_ready(&client, &base).await;
let entries: Vec<Value> = (0..51).map(|i| serde_json::json!({"content": format!("Entry {}", i)})).collect();
let resp = client.call_tool("batch_store", serde_json::json!({
"entries": entries
})).await;
assert!(resp.is_err() || resp.as_ref().unwrap().is_error());
let response = call_jsonrpc(
&client,
&base,
json!({
"jsonrpc": "2.0",
"id": "batch-too-large-1",
"method": "tools/call",
"params": {
"name": "batch_store",
"arguments": {
"entries": entries
}
}
}),
)
.await;
assert!(response.get("error").is_some(), "oversized batch_store should return an error");
Ok(())
}
#[tokio::test]
async fn e2e_batch_store_missing_content() -> anyhow::Result<()> {
let resp = client.call_tool("batch_store", serde_json::json!({
"entries": [{"content": "Valid entry"}, {"metadata": {}}]
})).await;
assert!(resp.is_err() || resp.as_ref().unwrap().is_error());
let base = base_url();
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(20))
.build()
.expect("reqwest client");
wait_until_ready(&client, &base).await;
let response = call_jsonrpc(
&client,
&base,
json!({
"jsonrpc": "2.0",
"id": "batch-missing-content-1",
"method": "tools/call",
"params": {
"name": "batch_store",
"arguments": {
"entries": [{"content": "Valid entry"}, {"metadata": {}}]
}
}
}),
)
.await;
assert!(response.get("error").is_some(), "missing batch entry content should return an error");
Ok(())
}
#[tokio::test]
async fn e2e_batch_store_appears_in_tools() -> anyhow::Result<()> {
let tools = client.list_tools().await?;
let parsed: Value = serde_json::from_str(&tools.content[0].text)?;
let names: Vec<&str> = parsed.as_array().unwrap().iter()
.filter_map(|t| t.get("name").and_then(|n| n.as_str()))
let base = base_url();
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(20))
.build()
.expect("reqwest client");
wait_until_ready(&client, &base).await;
let response = call_jsonrpc(
&client,
&base,
json!({
"jsonrpc": "2.0",
"id": "batch-tools-list-1",
"method": "tools/list",
"params": {}
}),
)
.await;
let names: Vec<&str> = response
.get("result")
.and_then(|value| value.get("tools"))
.and_then(Value::as_array)
.expect("tools/list result.tools")
.iter()
.filter_map(|t| t.get("name").and_then(Value::as_str))
.collect();
assert!(names.contains(&"batch_store"));
Ok(())
@@ -940,14 +1035,23 @@ async fn e2e_batch_store_appears_in_tools() -> anyhow::Result<()> {
#[tokio::test]
async fn e2e_existing_store_unchanged() -> anyhow::Result<()> {
let base = base_url();
let client = reqwest::Client::builder()
.timeout(Duration::from_secs(20))
.build()
.expect("reqwest client");
ensure_schema().await;
wait_until_ready(&client, &base).await;
let agent = format!("compat_{}", uuid::Uuid::new_v4());
let _ = db.purge_memories(&agent, None).await;
let resp = client.call_tool("store", serde_json::json!({
"agent_id": agent.clone(),
let _ = call_tool(&client, &base, "purge", json!({ "agent_id": agent, "confirm": true })).await;
let result = call_tool(&client, &base, "store", serde_json::json!({
"agent_id": agent,
"content": "Original store still works"
})).await?;
let result: Value = serde_json::from_str(&resp.content[0].text)?;
})).await;
assert!(result["success"].as_bool().unwrap_or(false));
db.purge_memories(&agent, None).await?;
Ok(())
}