Run the monitoring agent: $ARGUMENTS

Parse the mode from arguments:

If empty or "baseline": Run monitor agent with prompt "Baseline mode. Capture the current state of all in-scope targets. Use recon/ data if available for a richer baseline."
If "check": Run monitor agent with prompt "Check mode. Compare current state against baselines in monitor/. Report all changes and recommend which agents to re-run."
If "scope": Run monitor agent with prompt "Scope mode. Use MCP get_program_scope to check if the platform has updated the program scope. Diff against local scope.yaml and report new assets."

After the agent completes:

If changes were found, show a summary and suggest next actions
If scope changes were found, ask if the user wants to run /pipeline on new assets
Log to brain: uv run python3 $CLAUDE_PROJECT_DIR/tools/brain.py log "Monitor {mode}: {summary}"

Top-Tier Monitoring Bar

Treat changes as vulnerability triggers.