mr-memory 2.9.12 → 2.10.0

package/index.ts

@@ -15,11 +15,21 @@ import type { OpenClawPluginApi } from "openclaw/plugin-sdk";
 
 const DEFAULT_ENDPOINT = "https://api.memoryrouter.ai";
 
-/** Strip media-attached references so OpenClaw's image detection doesn't
- *  load old screenshots/photos from file paths embedded in memories. */
-const MEDIA_ATTACHED_RE = /\[media attached:\s*[^\]]*\]/gi;
+/** Strip media-attached references and OpenClaw media instruction text so
+ *  old screenshots/photos/audio don't get stored and re-injected forever. */
+const MEDIA_ATTACHED_RE = /\[media attached[^\]]*\]/gi;
+const MEDIA_TAGS_RE = /<media:(?:audio|image|video|document)>/gi;
+const MEDIA_INSTRUCTION_RE = /To send an image back,[\s\S]*?Keep caption in the text body\./gi;
+const MEDIA_INBOUND_PATH_RE = /\/Users\/[^\s]*\/media\/inbound\/[^\s)\]"]*/gi;
+const IMAGE_DATA_RE = /\[image data removed[^\]]*\]/gi;
 function stripMediaReferences(text: string): string {
-  return text.replace(MEDIA_ATTACHED_RE, "[media reference removed]");
+  return text
+    .replace(MEDIA_INSTRUCTION_RE, "")
+    .replace(MEDIA_ATTACHED_RE, "[media reference removed]")
+    .replace(MEDIA_TAGS_RE, "")
+    .replace(MEDIA_INBOUND_PATH_RE, "[media reference removed]")
+    .replace(IMAGE_DATA_RE, "")
+    .replace(/\n{3,}/g, "\n\n");
 }
 
 /** Wrap raw memory context in XML tags with a strong instruction */
@@ -38,6 +48,63 @@ function stripOldMemory(text: string): string {
   return text.replace(MEMORY_TAG_RE, "").replace(LEGACY_TAG_RE, "").trim();
 }
 
+// ── Ingest sanitization: strip system noise so only real conversations get stored
+// Patterns that indicate system-generated messages (not user conversations)
+const SYSTEM_NOISE_PATTERNS = [
+  // Heartbeat prompts and responses
+  /^Read HEARTBEAT\.md if it exists/,
+  /^HEARTBEAT_OK\s*$/,
+  // Pre-compaction flush prompts
+  /^Pre-compaction memory flush\./,
+  // Post-compaction audit warnings
+  /⚠️ Post-Compaction Audit:/,
+  // Silent reply ack
+  /^NO_REPLY\s*$/,
+  // Agent abort notices
+  /^Note: The previous agent run was aborted/,
+  // Queued message headers
+  /^\[Queued messages while agent was busy\]/,
+  // Media reference placeholders
+  /^\[media reference removed\]\s*$/,
+  // System event timestamps (cron fires, exec failures, etc.)
+  /^System: \[\d{4}-\d{2}-\d{2}/,
+  /^\[\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2} [A-Z]{2,4}\] ⚠️/,
+  // Memory flush with date pattern
+  /^Store durable memories now/,
+  // Current time system lines (injected by OpenClaw into system events)
+  /^Current time: .+\(America\//,
+];
+
+/** Strip OpenClaw envelope metadata from user messages */
+const ENVELOPE_METADATA_RE = /Conversation info \(untrusted metadata\):\s*```json\s*\{[^}]*\}\s*```\s*/g;
+const SENDER_METADATA_RE = /Sender \(untrusted metadata\):\s*```json\s*\{[^}]*\}\s*```\s*/g;
+const REPLY_CONTEXT_RE = /Replied message \(untrusted, for context\):\s*```json\s*\{[^}]*\}\s*```\s*/g;
+const MEMORY_INSTRUCTION_RE = /The above are retrieved memories from past conversations[^\n]*\n*/g;
+const IMPORTANT_MEMORY_RE = /IMPORTANT: The above (?:are retrieved memories|block contains retrieved memories)[^\n]*\n*/g;
+
+/** Sanitize text before storing in memory — removes system noise and envelope metadata */
+function sanitizeForIngest(text: string): string {
+  // 1. Strip our own memory injection wrapper (already exists as stripOldMemory)
+  let cleaned = stripOldMemory(text);
+  // 2. Strip envelope metadata blocks
+  cleaned = cleaned
+    .replace(ENVELOPE_METADATA_RE, "")
+    .replace(SENDER_METADATA_RE, "")
+    .replace(REPLY_CONTEXT_RE, "")
+    .replace(MEMORY_INSTRUCTION_RE, "")
+    .replace(IMPORTANT_MEMORY_RE, "");
+  // 3. Collapse excessive whitespace from removals
+  cleaned = cleaned.replace(/\n{3,}/g, "\n\n").trim();
+  return cleaned;
+}
+
+/** Check if a message is system noise that should not be stored */
+function isSystemNoise(text: string): boolean {
+  const trimmed = text.trim();
+  if (!trimmed) return true;
+  return SYSTEM_NOISE_PATTERNS.some(pattern => pattern.test(trimmed));
+}
+
 // Workspace files OpenClaw loads into the system prompt
 const WORKSPACE_FILES = [
   "IDENTITY.md", "USER.md", "MEMORY.md", "HEARTBEAT.md",
@@ -403,11 +470,17 @@ const memoryRouterPlugin = {
 
           const toStore: Array<{ role: string; content: string }> = [];
 
-          // Add the user message
+          // Add the user message (sanitized)
           if (lastUserIdx >= 0) {
             const userMsg = msgs[lastUserIdx] as { content?: unknown };
-            const userText = extractText(userMsg.content);
-            if (userText) toStore.push({ role: "user", content: stripMediaReferences(userText) });
+            let userText = extractText(userMsg.content);
+            if (userText) {
+              // Sanitize: strip memory injections, envelope metadata, and system noise
+              userText = sanitizeForIngest(stripMediaReferences(userText));
+              if (userText && !isSystemNoise(userText)) {
+                toStore.push({ role: "user", content: userText });
+              }
+            }
           }
 
           // Collect ALL assistant messages after the last user message
@@ -421,10 +494,18 @@ const memoryRouterPlugin = {
           }
           if (assistantParts.length > 0) {
             // Strip media refs so image paths don't get stored and re-injected
-            toStore.push({ role: "assistant", content: stripMediaReferences(assistantParts.join("\n\n")) });
+            let assistantText = stripMediaReferences(assistantParts.join("\n\n"));
+            // Skip storing system noise responses (HEARTBEAT_OK, NO_REPLY, etc.)
+            if (!isSystemNoise(assistantText)) {
+              toStore.push({ role: "assistant", content: assistantText });
+            }
           }
 
-          if (toStore.length === 0) return;
+          // Skip if nothing meaningful to store (both messages were system noise)
+          if (toStore.length === 0) {
+            log("memoryrouter: skipped ingest — system noise filtered out");
+            return;
+          }
 
           // Await the fetch so OpenClaw's runVoidHook keeps the event loop alive.
           try {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mr-memory",
-  "version": "2.9.12",
+  "version": "2.10.0",
   "description": "MemoryRouter persistent memory plugin for OpenClaw — your AI remembers every conversation",
   "type": "module",
   "files": [