[BUG] Tool server starts successfully but no endpoints respond (including /health) – health check always fails in v0.1.12

[munirohitrishi]

Title:
Tool server starts successfully but no endpoints respond (including /health) – health check always fails in v0.1.12

Description:

Environment

• Image: ghcr.io/usestrix/strix-sandbox:0.1.12
• Docker version: (add yours, e.g. 27.3.1)
• Host OS: Linux (Kali/WSL or similar)
• Reproduction: 100% consistent both in normal docker run and manual shell

What I expected

• The tool server should start and respond to health checks at http://127.0.0.1:<port>/health with something like {"status":"healthy"}
• The container entrypoint should detect it as healthy and continue starting Caido

What actually happens

1. In normal container startup (docker run ... ghcr.io/usestrix/strix-sandbox:0.1.12):

   • Shows: Starting tool server...
   • Then immediately: ERROR: Tool server failed to become healthy
   • Shows empty /tmp/tool_server.log (or no meaningful content)

2. When launching manually inside the container (most revealing):

   cd /app
   . venv/bin/activate
   python -m strix.runtime.tool_server \
     --token="ZBKGIrULg8nPTWvLMQMeeS1CuQHlRsQrAz/BBeii3brvNTPzVAhaHqF9uO62c2AguFsE7ePebynF4E2OkJN0Lg==.tK1hfCQpqtQDvPER5hXYq2S4dIRN1iCTgIf/VIS7d+Y=.guest" \
     --host 0.0.0.0 \
     --port 8081 \
     --timeout 120