Adds a new CLI subcommand that wraps external content with security boundaries,
mirroring the existing wrapExternalContent() function but making it accessible
for skills and manual API calls.
Features:
- Wrap content from stdin or URL
- Detect suspicious patterns (prompt injection attempts)
- Output wrapped content with security warnings
- JSON output option with metadata
Usage:
echo '{"data": "test"}' | moltbot security wrap --stdin --source api
moltbot security wrap --url https://api.example.com/data
curl -s https://api.example.com | moltbot security wrap --stdin
This addresses the gap where the existing external content protection only
applies to inbound hooks, not to content agents fetch manually.
Closes#4329
