feat: configurable LLM warm-up timeout, IP/CIDR targets, resume previous run, and improved Windows Docker guidance

[Ayush3603]

Changes

• [configurable timeout] Warm-up now respects STRIX_LLM_TIMEOUT environment variable so local models (Ollama, LMStudio) don't timeout prematurely
• [IP/CIDR targets] StrixAgent now includes ip_address and ip_range targets in task description for network scanning
• [resume previous scan] --resume RUN_NAME loads targets from agent_runs/<RUN_NAME>/run_metadata.json; run metadata is now persisted automatically
• [Windows Docker guidance] Improved error messaging for Windows users with Docker Desktop/WSL integration tips

Files Changed

• strix/interface/main.py - Configurable warm-up timeout, resume support
• strix/agents/StrixAgent/strix_agent.py - IP address/range task inclusion
• strix/telemetry/tracer.py - Persist run metadata for resume
• strix/interface/utils.py - Windows Docker troubleshooting

Usage

• Set timeout: $env:STRIX_LLM_TIMEOUT=600 (PowerShell)
• Resume: strix --resume <RUN_NAME>
• IP scan: strix --target 192.168.1.10 or strix --target 192.168.1.0/24

Fixes #56, #57, #58