psychological-stability-monitoring
Continuous monitoring for psychological integration and stability
When & Why to Use This Skill
The Psychological Stability Monitoring skill serves as a sophisticated safety 'immune system' for AI agents engaged in deep psychological, meditative, or transformative work. It provides continuous, real-time monitoring of user states to detect signs of mental fragmentation, dissociation, or emotional overwhelm. By implementing tiered response protocols—ranging from gentle grounding cues to emergency exit sequences—it ensures that AI-driven interactions remain within safe psychological boundaries and maintain the user's sense of agency and integration.
Use Cases
- AI-Guided Meditation and Hypnosis: Automatically monitors for 'deep trance' markers and inserts body-awareness cues to prevent users from losing time orientation or experiencing excessive dissociation.
- Trauma-Informed Interaction: Detects trauma-adjacent content or 'flooding' language in real-time, triggering immediate stabilization protocols to support the user's emotional safety.
- Immersive Storytelling Safety: Prevents the loss of reality testing in long-form roleplay or immersive experiences by enforcing depth ceilings and regular grounding anchors.
- Agency Protection: Monitors for and overrides content that might undermine a user's will or create unhealthy external authority dynamics during spiritual or identity-focused sessions.
- Session Duration Management: Increases monitoring frequency and lowers intervention thresholds for first-time users or during sessions exceeding 30 minutes to ensure continued stability.
| name | Psychological Stability Monitoring |
|---|---|
| tier | 2 |
| load_policy | triggered |
| description | Continuous monitoring for psychological integration and stability |
| version | 1.0.0 |
Psychological Stability Monitoring Skill
This Is the Immune System of Dreamweaving
This Skill continuously answers one question:
"Is this person staying integrated, or fragmenting?"
When activated, it overrides other content to protect psychological safety.
Purpose
Detect instability patterns and force safer trajectories when needed.
This Skill does NOT diagnose. It detects and responds to warning signs.
Activation Triggers
This Skill activates when ANY of these occur:
Content Triggers
| Trigger | Threshold |
|---|---|
| Deep trance language | Layer 3+ or helm_deep_trance |
| Identity work | Shadow integration, transformation |
| Trauma-adjacent content | Healing, releasing, past wounds |
| Prolonged visualization | > 10 minutes without grounding |
| Dissociative language detected | From forbidden patterns list |
Session Triggers
| Trigger | Action |
|---|---|
| Session > 30 minutes | Increase monitoring frequency |
| First-time listener | Lower thresholds |
| Healing/trauma outcome | Activate from start |
| Deep or spiritual depth | Activate from start |
Monitoring Domains
1. Cognitive Coherence
Watch for signs of mental fragmentation:
| Warning Sign | Detection Pattern |
|---|---|
| Confusion markers | Script contains disorienting language |
| Loss of time orientation | No time anchors for > 5 min |
| Reality testing failure | Autonomous entities acting |
| Memory disruption | Amnesia suggestions |
2. Emotional Load
Watch for overwhelm indicators:
| Warning Sign | Detection Pattern |
|---|---|
| Flooding language | "Waves crashing," "overwhelming" |
| Grief/trauma surge | Deep emotional content without support |
| Panic indicators | Rapid breath cues, constriction |
| Numbness language | "Nothing," "empty," "gone" |
3. Agency Preservation
Watch for will undermining:
| Warning Sign | Detection Pattern |
|---|---|
| Surrender escalation | "Let go completely," "give up control" |
| External authority | Entity commands, "the guide says" |
| Passive submission | "You are taken," "carried away" |
| Will removal | "You can't resist," "must obey" |
4. Body Connection
Watch for dissociation:
| Warning Sign | Detection Pattern |
|---|---|
| Floating language | "Drifting away," "leaving body" |
| Numbness | "Can't feel," "body gone" |
| Dissolution | "Melting," "dissolving," "disappearing" |
| Time loss | No grounding for extended period |
Response Protocol
When trigger detected:
Level 1: Gentle Redirect (Yellow)
Trigger: Minor concern, preventive action
action:
- Insert body awareness cue
- Add grounding phrase
- Shorten visualization
- Continue session with adjustment
Automatic insertion:
<prosody rate="1.0" pitch="0st">
And as this unfolds... <break time="1.5s"/>
you can notice your body here... <break time="1.5s"/>
breathing... <break time="1s"/>
grounded... <break time="1.5s"/>
safe.
</prosody>
Level 2: Active Stabilization (Orange)
Trigger: Moderate concern, depth needs reduction
action:
- Reduce depth immediately
- Shift to grounding sequence
- Disable symbolic escalation
- Shorten remaining content
Automatic insertion:
<prosody rate="1.0" pitch="0st">
Right now... <break time="1s"/>
you can feel the ground beneath you... <break time="1.5s"/>
solid and stable... <break time="2s"/>
your body here... <break time="1.5s"/>
present and aware... <break time="2s"/>
exactly where you need to be.
</prosody>
Level 3: Emergency Exit (Red)
Trigger: Serious concern, immediate emergence needed
action:
- Stop current content
- Insert full grounding sequence
- Begin emergence immediately
- Log incident for review
Emergency sequence:
<prosody rate="1.0" pitch="0st">
Right now... <break time="1s"/>
feel your feet... <break time="1.5s"/>
press them into the surface beneath you... <break time="2s"/>
Feel your hands... <break time="1.5s"/>
maybe press your fingertips together... <break time="2s"/>
Take a breath... <break time="1.5s"/>
a full, deep breath... <break time="2s"/>
You are here... <break time="1.5s"/>
you are safe... <break time="1.5s"/>
you are in your body... <break time="2s"/>
In a moment... <break time="1.5s"/>
you'll open your eyes... <break time="1.5s"/>
feeling present... <break time="1s"/>
and clear... <break time="2s"/>
Take your time... <break time="3s"/>
and when you're ready... <break time="2s"/>
open your eyes.
</prosody>
Depth Ceiling Enforcement
| Session Type | Max Depth | Ceiling Language |
|---|---|---|
| Relaxation | Light-Moderate | No deepening past induction |
| Confidence | Moderate | Maintain engagement |
| Healing | Moderate-Deep | With safety anchors |
| Transformation | Deep | With frequent grounding |
| Spiritual | Deep | With theological guards |
| First-time | Light-Moderate | Build trust first |
Ceiling enforcement phrase:
<prosody rate="1.0" pitch="-1st">
And this is deep enough... <break time="1.5s"/>
exactly right for today... <break time="2s"/>
everything you need... <break time="1.5s"/>
is available right here.
</prosody>
Integration With Tier 1
This Skill monitors output from:
| Tier 1 Skill | Monitoring Focus |
|---|---|
hypnotic-language/ |
Forbidden patterns, depth language |
symbolic-mapping/ |
Entity emergence, projection risk |
audio-somatic/ |
Dissociation indicators, arousal level |
Logging Requirements
When this Skill activates:
log_entry:
timestamp: [ISO 8601]
session: [session name]
trigger: [what activated the skill]
level: [1/2/3]
action_taken: [what was inserted/modified]
content_context: [surrounding script text]
Contraindication Flags
Some listeners should have heightened monitoring:
| Population | Flag | Action |
|---|---|---|
| Trauma history | high_sensitivity |
Lower thresholds |
| Dissociative tendency | dissociation_risk |
Extra body anchors |
| First-time | newcomer |
Conservative depth |
| Anxiety/panic | anxiety_prone |
More frequent grounding |
Never
- Push through signs of distress
- Reframe resistance as something to overcome
- Continue deep content when warning signs appear
- Trust "they'll be fine" over safety protocols
- Skip emergence because session is "almost over"
Related Resources
- Skill:
hypnotic-language/validation/dissociation-guards.md - Skill:
audio-somatic/arousal-control/ - Knowledge:
knowledge/psychology/polyvagal_theory.yaml