fix: is_spam always returned None due to missing return statement

[Ridwannurudeen]

Summary

Fixes #4.

is_spam() in reply_manager.py built an extensive list of regex patterns for detecting crypto spam but never evaluated them — the function fell off the end and implicitly returned None. As a result, _should_reply() never applied the -3 spam penalty, and the agent replied to every piece of spam as if it were legitimate.

Root cause

# Before — no return statement; function silently returns None
def is_spam(self, content: str) -> bool:
    ...
    patterns = [ ... ]
    # <-- nothing here

Fix

# After
        return any(re.search(pattern, clean) for pattern in patterns)

Also removes two redundant per-call imports (import re and from unicodedata import normalize) that duplicated the module-level imports already present at the top of the file.

Changes

• agent/engines/twitter/reply_manager.py
  • Add missing return to is_spam()
  • Move from unicodedata import normalize to module-level imports
  • Remove import re / from unicodedata import normalize from inside the function body

Test plan

• is_spam("🚀 100x moon soon buy now") returns True
• is_spam("interesting take on the consensus mechanism") returns False
• _should_reply() correctly penalises spam content with a -3 score adjustment