mindgraph-mcp 0.25.0 → 0.26.0

package/README.md

@@ -97,8 +97,8 @@ makes: `POST /retrieve/context` with up to 8 keywords from the prompt (obvious
 secrets redacted first), 5 items, no source chunks — and injects at most 4,000
 characters headed `MindGraph recall (bounded):`, one line per item (type,
 label, one-line summary, source document). An item is injected only when its
-label or summary mentions two distinct keywords from the prompt (one, when the
-prompt yields a single keyword), so a prompt can inject nothing. Prompts under 12 characters and
+label or summary mentions two distinct keywords from the prompt, so a prompt
+can inject nothing (a prompt with a single usable keyword never injects). Prompts under 12 characters and
 slash commands are skipped, a block identical to the last one injected is not re-injected in the same
 session, and the hook gives up after 4 s and injects nothing on any error, so
 it never blocks a prompt. Turn it off with
@@ -165,6 +165,12 @@ checkouts)? Declare them once in `.mindgraph/workspace.json` at the parent:
   Anchoring requires an explicit space — per-repo `space_uid`, a
   `repo_space_uid` argument, or `MINDGRAPH_CODE_SPACE_UID` — so shared code
   identity never lands silently in a private per-agent space.
+- **Work brief**: a session started at the workspace root, or one folder
+  above it, selects durable work across every declared repository and lists
+  the other eligible tasks by repository. `mindgraph_memory` with
+  `mode: "resume"` and `scope: "workspace"` returns the same view on demand.
+  Only the brief and that tool look one folder down for a workspace file;
+  routing and sync use the upward search described above.
 - **Memory-file sync**: `mindgraph_sync action=scan workspace=true` fans out
   across the declared repositories (and errors with a diagnostic when nothing
   is declared); all other sync actions stay single-repository.