---

“I’m feeling anxious about a situation and need support. Can you help me check whether my thoughts are realistic, distorted, or possibly based in fear rather than fact?”

“I’m stuck in a spiral of anxious thoughts. Can you offer a grounded perspective that includes both possibilities and limitations of what might be going on?”

Encourages curiosity and ambiguity, reducing certainty-based distortions common in anxiety.

“I’m struggling with anxious thoughts that feel very real, but I know they might not be. Please give me feedback with the assumption that I might not be seeing things clearly.”

Why it matters: Explicitly names the possibility of impaired insight—a crucial step for those experiencing paranoia, delusion, or OCD-type spirals.

“Before I ask for advice, can you help me slow down by naming what might be happening in my body when I feel this anxious?”

Guardrail: Redirects attention to somatic cues, supporting nervous system regulation before cognitive processing. (Polyvagal-informed.)

“Here’s what I’m telling myself right now. Can you help me sort out what is emotion, what is sensation, and what might be the actual story?”

Cognitive integrity support: Teaches separation of feeling from fact, which is often collapsed in anxious states.

“Can you help me reflect on whether the question I’m asking comes from a grounded part of me—or an overwhelmed/anxious/distorted part?”

Built-in check: Encourages parts work or ego-state awareness. Prevents taking action based on the most activated voice.

“I’m afraid of something and want to talk about it, but I also want to make sure my fear isn’t driving the whole story. Can you give me both a supportive and a skeptical lens?”

Balance support: Prevents uncritical affirmation of fear-based thoughts, offers containment.

“Here’s the anxious thought I’m having. Can you help me rewrite it as if I were speaking from a calmer, wiser, or future version of myself?”

Adaptive narrative support: Encourages the use of internal resources rather than solely external validation.

“These thoughts feel intense and possibly unsafe. Before offering support, can you first check with me if I’m in a safe place and have a plan if things get worse?”

Ethical guardrail: Prioritizes real-world safety over emotional bypassing or reassurance loops.

“I’m having thoughts I’m not sure are based in reality. Please don’t affirm anything I say until you ask a few grounding questions to help assess whether this is a common anxiety pattern or something more concerning.”

Proactive protection: Prevents AI from unintentionally reinforcing delusional thinking or paranoia. This is especially helpful in early psychosis or trauma with dissociation.