mem-search

1# Smart Search (Default)
2 
3Search past work across all sessions. Use the original three-step mem-search workflow by default: search → timeline → selected get_observations. Reason about the index and context before choosing details. Do not use auto or guided modes.
4 
5## Runtime and Tool Names
6 
7The examples below use the local worker tools. Native Pi and DSH integrations
8name the same three steps `mem_search`, `mem_timeline`, and
9`mem_get_observations`. Native OpenCode uses `claude_mem_search`,
10`claude_mem_timeline`, and `claude_mem_get_observations`. CMEM Pro uses `memory_search`, `timeline`, and
11`get_observations`; use the parameter names advertised by that connection.
12 
13A self-hosted server runtime uses the same three steps: `search`, `timeline`,
14and `get_observations`, scoped to the configured server project. Server
15observation IDs are strings. Set `projectId` only when selecting another
16authorized server project. Dumb search is unavailable in server runtime. A direct self-hosted `/v1/mcp`
17connection instead requires `projectId` on each call and `anchorId` for its
18timeline. CMEM Pro also uses `anchorId` for timeline context, and uses
19`get_summaries` when the selected index rows are session summaries. Follow the
20active connection’s advertised schema.
21 
22## When to Use
23 
24Use when users ask about PREVIOUS sessions (not current conversation):
25 
26- "Did we already fix this?"
27- "How did we solve X last time?"
28- "What happened last week?"
29 
30## Layered Workflow (ALWAYS Follow)
31…