Internal-state triggers are high-risk — replace…

When This Fires

• When designing triggers and the best candidate is an internal state (see Trigger stress agents on body signals… for stress)
• When an internal-state trigger isn't firing reliably — low interoceptive accuracy may be the cause
• When deciding between investing in interoceptive calibration vs. replacing with an external trigger
• During trigger audits when assessing reliability risk across your agent portfolio

Common Failure Mode

Assuming you can detect internal states because you can label them retrospectively: "I know when I'm stressed because I remember being stressed yesterday." Retrospective labeling uses different cognitive pathways than real-time detection. You need to detect the state as it's occurring, not after it's resolved. Most people have much lower real-time interoceptive accuracy than they believe.

The Protocol

(1) When a trigger depends on internal state detection, flag it as high-risk. (2) Option A — Replace: find the nearest external observable event that correlates with the internal state. "When I feel distracted" → "After the third time I switch apps in 5 minutes." "When my energy drops" → "At 2:30 PM" (a time when energy reliably drops). (3) Option B — Calibrate: invest in deliberate interoceptive training for this specific state. Practice detection during known instances. Verify accuracy against external measures. Budget 2-4 weeks of practice before trusting the trigger operationally. (4) Default to Option A unless the internal state has no reliable external proxy and you're willing to invest in calibration. (5) Even after calibration, monitor the trigger's firing rate (Define agent success as 80%+ firing rate, not subjective…) — interoceptive accuracy fluctuates with fatigue, stress, and attention.