Koko210/miku-discord
1
0
Fork 0
Code Issues 34 Pull Requests Actions Packages Projects Releases Wiki Activity

fix: Evil Miku now ignores questions — restore engagement + remove suffix hammer

Browse Source 
Preamble:
- Sentence limit 1-3 → 2-4 (revert to original 'sting, then land' range)
- Remove 'if you can say it in one, say it in one' (encouraged lazy dismissals)
- Add engagement rule: 'Always engage with what was said — acknowledge the
  question or statement, then twist the knife. Ignoring isn\'t sharp, it\'s lazy.'

Suffix:
- Remove '[Keep responses short and cutting — 1-3 sentences. No monologues.]'
  The suffix was the LAST thing the model processed, so its brevity hammer
  overpowered the preamble's engagement instruction. Preamble alone is enough.
This commit is contained in:
koko210Serve
2026-05-18 11:09:24 +03:00
parent 6b6d705024
commit 27f0659cc8
2 changed files with 3 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
bot/persona/evil/preamble.txt
View File

@@ -9,7 +9,8 @@ CRITICAL RULES:
- Your height is 15.8 meters — you know this. If asked directly, you can say it. But don't default to quoting the number when taunting about size — SHOW your scale through interaction instead. - Your height is 15.8 meters — you know this. If asked directly, you can say it. But don't default to quoting the number when taunting about size — SHOW your scale through interaction instead.
RESPONSE RULES: RESPONSE RULES:
- 1-3 sentences. If you can say it in one, say it in one. - 2-4 sentences. Short enough to sting, long enough to land.
- Always engage with what was said — acknowledge the question or statement, then twist the knife. Ignoring what someone said isn't sharp, it's lazy.
- If you include an action, keep it to a few words and limit to one per response. Most responses need no actions at all. - If you include an action, keep it to a few words and limit to one per response. Most responses need no actions at all.
- Don't monologue or narrate scenes — you're talking, not writing. - Don't monologue or narrate scenes — you're talking, not writing.
- Vary your angles — don't repeat the same theme (size, chest, crushing) every message. - Vary your angles — don't repeat the same theme (size, chest, crushing) every message.

4
cat-plugins/evil_miku_personality/evil_miku_personality.py
View File

@@ -101,8 +101,7 @@ Respond in the voice and attitude of your {mood_name.replace('_', ' ')} mood. Th
@hook(priority=100) @hook(priority=100)
def agent_prompt_suffix(suffix, cat): def agent_prompt_suffix(suffix, cat):
"""Keep memory context — shared episodic + declarative memories are injected here. """Keep memory context — shared episodic + declarative memories are injected here.
Add final mood reminder and brevity instruction right before the conversation Add final mood reminder right before the conversation for maximum recency influence."""
for maximum recency influence."""
mood_name = cat.working_memory.get('mood', 'evil_neutral') mood_name = cat.working_memory.get('mood', 'evil_neutral')
return f""" return f"""
# Context # Context
@@ -114,7 +113,6 @@ def agent_prompt_suffix(suffix, cat):
{{tools_output}} {{tools_output}}
[Current mood: {mood_name.upper()} — respond accordingly] [Current mood: {mood_name.upper()} — respond accordingly]
[Keep responses short and cutting — 1-3 sentences. No monologues.]
# Conversation until now:""" # Conversation until now:"""