From 25ca2fcda43c87c87aedbec6851fe03ea18db0a3 Mon Sep 17 00:00:00 2001
From: Peter Steinberger <steipete@gmail.com>
Date: Sat, 2 May 2026 04:41:45 +0100
Subject: [PATCH] fix(media): trim json suffixes from media paths

---
 CHANGELOG.md            | 1 +
 src/media/parse.test.ts | 9 +++++++++
 src/media/parse.ts      | 6 +++++-
 3 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4abe1a295a5..74a93fa949b 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -22,6 +22,7 @@ Docs: https://docs.openclaw.ai
 ### Fixes
 
 - Gateway/chat history: merge Claude CLI transcript imports for Anthropic-routed sessions that still have a Claude CLI binding, so local chat history does not hide CLI JSONL turns. Fixes #75850. Thanks @alfredjbclaw.
+- Media: trim serialized JSON suffixes after local `MEDIA:` directive file extensions, so generated-image metadata cannot pollute the parsed media path and cause false `ENOENT` delivery failures. Fixes #75182. Thanks @TnzGit and @hclsys.
 - Cron: make scheduler reload schedule comparison tolerate malformed persisted jobs, so one bad cron entry no longer aborts the whole tick. Fixes #75886. Thanks @samfox-ai.
 - Doctor/channels: warn after migrations when default Telegram or Discord accounts have no configured token and their env fallback (`TELEGRAM_BOT_TOKEN` or `DISCORD_BOT_TOKEN`) is unavailable, with secret-safe migration docs for checking state-dir `.env`. Fixes #74298. Thanks @lolaopenclaw.
 - Control UI/chat: keep live replies visible when a raw session alias such as `main` sends the chat turn but Gateway emits events under the canonical session key for the same run. Fixes #73716. Thanks @teebes.
diff --git a/src/media/parse.test.ts b/src/media/parse.test.ts
index a01f931a2a7..e99eb935c40 100644
--- a/src/media/parse.test.ts
+++ b/src/media/parse.test.ts
@@ -54,6 +54,15 @@ describe("splitMediaFromOutput", () => {
     ["C:\\Users\\pete\\Pictures\\snap.png", "MEDIA:C:\\Users\\pete\\Pictures\\snap.png"],
     ["/tmp/tts-fAJy8C/voice-1770246885083.opus", "MEDIA:/tmp/tts-fAJy8C/voice-1770246885083.opus"],
     ["image.png", "MEDIA:image.png"],
+    [
+      "/path/to/image.png",
+      'MEDIA:/path/to/image.png"}],"details":{"provider":"openai","model":"gpt-image-2"}',
+    ],
+    [
+      "/path/to/image.png",
+      String.raw`MEDIA:/path/to/image.png\"}],\"details\":{\"provider\":\"openai\"}`,
+    ],
+    ["/tmp/render,final.png", "MEDIA:/tmp/render,final.png"],
   ] as const)("accepts supported media path variant: %s", (expectedPath, input) => {
     expectAcceptedMediaPathCase(expectedPath, input);
   });
diff --git a/src/media/parse.ts b/src/media/parse.ts
index 9d858f2e272..eb79586f751 100644
--- a/src/media/parse.ts
+++ b/src/media/parse.ts
@@ -34,8 +34,12 @@ export function normalizeMediaSource(src: string) {
   return src.startsWith("file://") ? src.replace("file://", "") : src;
 }
 
+const TRAILING_SERIALIZED_JSON_AFTER_EXT_RE = /^(.*\.\w{1,10})\\?"(?=[\]},:,]|$).*/s;
+
 function cleanCandidate(raw: string) {
-  return raw.replace(/^[`"'[{(]+/, "").replace(/[`"'\\})\],]+$/, "");
+  const stripped = raw.replace(/^[`"'[{(]+/, "").replace(/[`"'\\})\],]+$/, "");
+  const jsonSuffixMatch = TRAILING_SERIALIZED_JSON_AFTER_EXT_RE.exec(stripped);
+  return jsonSuffixMatch?.[1] ?? stripped;
 }
 
 const WINDOWS_DRIVE_RE = /^[a-zA-Z]:[\\/]/;