mcp-agents-memory 0.6.0 → 0.6.2

package/README.md

@@ -90,12 +90,18 @@ Claude Desktop / Claude Code / any MCP-aware client:
 {
   "mcpServers": {
     "memory": {
-      "command": "mcp-agents-memory"
+      "command": "mcp-agents-memory",
+      "env": {
+        "AGENT_KEY": "agent_claude",
+        "AGENT_PLATFORM": "claude-code"
+      }
     }
   }
 }
 ```
 
+`AGENT_PLATFORM` is recorded as the Curator's harness identity on every memory_add call. The Curator's model is captured per-call (defaulting to the Producer's author_model) — set explicitly via the curator_model argument when an orchestrator saves memories on behalf of a different model (e.g. delegating to a subagent). This avoids the staleness that env-static model values would introduce when /model swaps mid-session.
+
 ### Cross-machine memory
 
 On a second computer, run `npm i -g mcp-agents-memory` and `mcp-agents-memory setup` pointing to the **same** `DATABASE_URL`. Memory shares automatically — the database is the source of truth and the MCP server is stateless.

package/build/index.js

@@ -116229,9 +116229,10 @@ async function processBatch(text, subjectId, projectId, rawSource, provenance =
            subject_id, project_subject_id, content, fact_type,
            confidence, importance, tags, embedding, validation_status,
            author_model, platform, session_id,
+           agent_platform, agent_model,
            author_model_id, platform_id, effective_confidence, source
          )
-         VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16)
+         VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15, $16, $17, $18)
          RETURNING id`,
         [
           subjectId,
@@ -116246,6 +116247,8 @@ async function processBatch(text, subjectId, projectId, rawSource, provenance =
           provenance.author_model,
           provenance.platform,
           provenance.session_id,
+          provenance.agent_platform,
+          provenance.agent_model,
           resolvedModel?.id ?? null,
           resolvedPlatform?.id ?? null,
           effectiveConfidence,
@@ -117507,24 +117510,31 @@ function registerTools(server) {
         subject_key: external_exports.string().optional().default(process.env.MEMORY_DEFAULT_SUBJECT || "default_user").describe("Primary subject key."),
         project_key: external_exports.string().optional().describe("Project key if relevant."),
         author_model: external_exports.string().optional().describe("Model name or alias (e.g., 'sonnet', 'opus', 'gemini')."),
+        curator_model: external_exports.string().optional().describe("Curator model override \u2014 the model running memory_add. Defaults to author_model (Producer == Curator solo case)."),
         platform: external_exports.string().optional().describe("Platform (e.g., 'antigravity', 'claude-code')."),
         session_id: external_exports.string().optional().describe("Unique session identifier.")
       }
     },
     async (args) => {
+      const agentPlatform = process.env.AGENT_PLATFORM;
       const subjectId = await getOrCreateSubject(args.subject_key, "person");
       let projectId = null;
       if (args.project_key) {
         projectId = await getOrCreateSubject(args.project_key, "project");
       }
+      const authorModel = args.author_model ?? null;
+      const curatorModel = args.curator_model ?? args.author_model ?? null;
+      const platform = args.platform ?? agentPlatform;
       const result = await processBatch(
         args.text,
         subjectId,
         projectId,
         args.text,
         {
-          author_model: args.author_model,
-          platform: args.platform,
+          author_model: authorModel,
+          platform,
+          agent_platform: agentPlatform,
+          agent_model: curatorModel,
           session_id: args.session_id
         }
       );