The Codex MCP tool description was misleading - it said "Ask a question"
and "Returns a thoughtful response" which implied it was read-only.
In reality, Codex has full filesystem access and can create/modify files
directly since the MCP server executes the `codex` CLI agent which has
full implementation capabilities.
**Changes:**
- Updated description to clarify Codex can "implement changes or answer questions"
- Added note that "Codex has full filesystem access and can create/modify files directly"
- Updated inputSchema description from "question" to "question or implementation task"
- Updated comment from "for initial questions" to "for both questions and implementation"
This prevents confusion about Codex being a read-only Q&A tool when it's
actually a full implementation agent.
Addresses the root cause of issue #631 (infinite Docker agent restart loop)
and prevents similar issues with host-agent and sensor-proxy.
Changes:
- Set dockeragent.Version default to "dev" instead of hardcoded version
- Add version embedding to server build in Dockerfile
- Add version embedding to host-agent builds (all platforms)
- Add version embedding to sensor-proxy builds (all platforms)
This ensures:
1. Server's /api/agent/version endpoint returns correct v4.26.0
2. Downloaded agent binaries have matching embedded versions
3. Dev builds skip auto-update (Version="dev")
4. No version mismatch triggers infinite restart loops
Related to #631
