fix(tool): fix bug where os-vision tools wrongly use path from sandbox for reading images. (#64)

* upd: add futurex evaluation support.

* upd: support multiple eval for futurex and add relavent doc.

* upd: fix bugs with doc for futurex.

* debug: fix wrong calling path.

* add preparation for finsearchcomp.

* update a premature version of finsearchcomp benchmark.

* clean redundent code in merging.

* upd: modify yaml to use Mirothinker as the main agent, add check progress file to exclude T1.

* upd: check_progress function for finsearchcomp now consider globe and greater china respectively.

* upd: add docs and shell script for multiple runs.

* fix: check_finsearchcomp_progress not displaying results from greater china region.

* fix: catch ContextLimitError in more observed cases.

* initialize open source tools for audio, vision and reasoning.

* upd: docs for open-source tools.

* fix wrong date.

* fix bug with os vision_mcp_server loading images from sandbox paths.

Author: JubSteven

src/tool/mcp_servers/vision_mcp_server_os.py

@@ -92,6 +92,8 @@ async def visual_question_answering(image_path_or_url: str, question: str) -> st
                         )
                     else:
                         return f"Failed to fetch image from URL: {image_path_or_url}"
+        elif "home/user" in image_path_or_url:
+            return "The visual_question_answering tool cannot access to sandbox file, please use the local path provided by original instruction"
         else:
             messages_for_llm[0]["content"][0]["image_url"]["url"] = image_path_or_url