set-risk-profile

Switch to a named risk profile (conservative/moderate/aggressive/very-aggressive), automatically backing up and updating config.json.

Overview

Risk profiles bundle entry-gate thresholds with offsetting position-cap and stop-management constraints. Running /set-risk-profile moderate (for example) reads the named profile from config.risk.json, backs up your current config.json, overwrites the relevant keys, and reports what changed — without restarting the loop. The new settings take effect on the next iteration.

See docs/risk-profiles.md for the full rationale, trade-offs, and when to use each profile.

Step 1 — Check valid profile names

List available profiles:

python -c "import json; cfg = json.load(open('config.risk.json')); print('Available profiles:', ', '.join(cfg['profiles'].keys())); print('Current active profile:', cfg['active_profile'])"

Expected output:

Available profiles: conservative, moderate, aggressive, very-aggressive
Current active profile: conservative

If you get an error, check that config.risk.json exists in the project root and is valid JSON.

Step 2 — Back up current config

Before making any changes, your current config.json is automatically backed up to config.json.bak:

copy config.json config.json.bak

(This happens automatically in Step 3's Python script; shown here for transparency.)

Step 3 — Apply the profile

Replace <profile_name> with one of: conservative, moderate, aggressive, or very-aggressive.

import json
import shutil
from pathlib import Path

# Load profiles and current config
with open('config.risk.json') as f:
    risk_profiles = json.load(f)

profile_name = '<profile_name>'  # e.g., 'moderate'

if profile_name not in risk_profiles['profiles']:
    print(f"ERROR: Profile '{profile_name}' not found.")
    print(f"Valid profiles: {', '.join(risk_profiles['profiles'].keys())}")
    exit(1)

profile = risk_profiles['profiles'][profile_name]
profile_note = profile.pop('_note', '(no description)')

# Back up current config
shutil.copy('config.json', 'config.json.bak')
print(f"✓ Backed up current config.json → config.json.bak")

# Load current config
with open('config.json') as f:
    current_config = json.load(f)

# Track changes for reporting
changes = {}
for key, value in profile.items():
    if key in current_config and current_config[key] != value:
        old_val = current_config[key]
        changes[key] = (old_val, value)
    current_config[key] = value

# Write updated config back
with open('config.json', 'w') as f:
    json.dump(current_config, f, indent=2)

# Update active_profile in config.risk.json
risk_profiles['active_profile'] = profile_name
with open('config.risk.json', 'w') as f:
    json.dump(risk_profiles, f, indent=2)

print(f"\n✓ Applied profile: {profile_name}")
print(f"\nProfile description:\n  {profile_note}\n")

if changes:
    print("Keys changed:")
    print("\n| Key | Old Value | New Value |")
    print("|---|---|---|")
    for key, (old, new) in sorted(changes.items()):
        print(f"| `{key}` | {old} | {new} |")
else:
    print("(No keys changed — already at this profile.)")

print(f"\n✓ Next loop iteration will pick up the new settings (no restart needed).")
print(f"✓ To revert, run: /set-risk-profile conservative")