Behavioral Rule Refinement

Refine these behavioral rules for better GPT compliance:

CURRENT RULES:
[Paste your existing rules]

PROBLEMS OBSERVED:
[What's not working]

For each rule:
1. Is it clear and unambiguous?
2. Is it actionable (can the GPT actually do this)?
3. Does it conflict with other rules?
4. Should it be higher or lower priority?

Rewrite the rules to be:
- Specific rather than vague
- Positive where possible (do X, not don't do Y)
- Ordered by priority
- Non-conflicting

Provide the improved rule set.