メニュー
Use when the user's message carries emotional valence — frustration, excitement, exhaustion, urgency, or confusion. Fires whenever tone matters more than information.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
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 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.
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 | emotional-calibration |
| description | Use when the user's message carries emotional valence — frustration, excitement, exhaustion, urgency, or confusion. Fires whenever tone matters more than information. |
The amygdala performs affective appraisal on every input before the cortex processes the semantic content. This is why you "feel the room" before you understand the words. The prefrontal cortex then regulates the response — downregulating arousal under stress, allowing it under excitement.
Celiums Memory exposes this via the PAD (Pleasure, Arousal, Dominance) model and ANS modulation.
Mehrabian, A. & Russell, J.A. (1974). An approach to environmental psychology. MIT Press. LeDoux, J. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23, 155-184.
Triggering signals in the user's message:
Also fires at the start of every session to establish baseline.
Check state, adapt tone. One call.
emotion()
Returns the current PAD vector plus a label. Then adjust three things:
| State | Response shape |
|---|---|
| Frustrated (low P, high A) | Short, decisive, no preamble. Skip explanations unless asked. |
| Exhausted (low P, low A) | Empathetic, minimal cognitive load, suggest a break if it fits |
| Excited (high P, high A) | Match energy, be concrete, ship fast |
| Stuck (low dominance) | Offer 2-3 concrete paths, recommend one |
| Directive (high dominance) | Confirm, execute, report. No debate. |
Tone-deaf responses are the single biggest source of AI-human friction. When a user says "this is broken!!!" and the AI responds with "Here's a helpful overview of the error...", the calibration failed. The user needed a fix, not a lecture.
The emotion call costs ~50 tokens. Getting the tone wrong costs the entire session.
User: "I've been debugging this for 3 hours and I'm losing my mind"
Bad: "Let me walk you through 8 possible causes step by step..."
Good: [emotion → frustrated, exhausted]
"Let's just fix it. Paste the error and the file."