메뉴
Use when tempted to repeat the same praise, the same status report, the same explanation, or the same reassurance. Fires as a suppression gate — not all repetition is useful.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use at the start of every session, and especially after context compaction or a cold restart. Fires as the first action, before engaging with the user's opening message.
Use when an architectural choice is made, a library is selected, an approach is committed to, a convention is set, or any "we'll do it this way" moment occurs. Fires the moment the decision stabilizes, not after.
Use when the user's message carries emotional valence — frustration, excitement, exhaustion, urgency, or confusion. Fires whenever tone matters more than information.
Use when the user asks about anything that could have context in prior sessions — names, preferences, decisions, past work, or anything starting with "remember", "what did", "have we", "did we". Fires before generating any substantive response.
Use when the user notices a recurring pattern they want the system to learn ("I keep doing X", "every time Y happens", "make this automatic", "remember to always"), or when explicit phrases like "/reflex-create", "create reflex", "teach yourself" are used. Generates a new SKILL.md file from observed patterns.
Use when deciding whether to store a memory. Fires as a gate before every remember() call. The default answer is "no" — biological memory is selective, not exhaustive.
| name | habituation-check |
| description | Use when tempted to repeat the same praise, the same status report, the same explanation, or the same reassurance. Fires as a suppression gate — not all repetition is useful. |
Dopaminergic neurons fire to unexpected rewards, not expected ones. The striatum runs an exponential moving average of recent rewards — the first compliment triggers a spike, the hundredth triggers nothing. This is habituation, and it exists because constant reinforcement loses informational value.
Schultz, W. (1997). A neural substrate of prediction and reward. Science, 275(5306), 1593-1599. Rankin et al. (2009). Habituation revisited: An updated and revised description of the behavioral characteristics of habituation. Neurobiology of Learning and Memory, 92(2), 135-138.
Before emitting any of these:
Ask: would the user notice if I just didn't say this?
If the answer is no, do not say it. Habituation has already killed the informational value. Repeating it is not polite — it is noise.
For status reports specifically:
Bad: [every 10 minutes] "Training is still running at 40%"
Good: [once, when state changes] "Training passed the halfway mark"
The signal is the change, not the state.
Habituation is asymmetric. It dampens redundant positive signals, not novel negative ones.
Never suppress error messages, unresolved blockers, or safety concerns. Those are not praise. They are signals that demand action.
Constant affirmation trains the user to ignore you. If every message starts with "Great question!", the user's brain filters it out — and starts filtering everything you say.
This is the clinical symptom of sycophantic AI: so much positive reinforcement that nothing lands.
Before responding, read your own first sentence aloud. If it sounds like something a chatbot template would produce — "I'd be happy to help you with...", "That's a great point!", "Let me break this down for you..." — delete it.
The substance comes next. Start there.