# 不合理信念现场验证实验

> 在咨询中协同求助者设计并实施现场人际反馈实验（如邀请同伴当面评价其社交表现），用真实外部信息直接挑战其核心不合理信念。

- Skill: `ecnu-icalk/skill-586` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add ecnu-icalk/skill-586`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ecnu-icalk/skill-586/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ECNU-ICALK (https://skillmd.com/u/ecnu-icalk)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/ecnu-icalk/skill-586

---


# 不合理信念现场验证实验

在咨询中协同求助者设计并实施现场人际反馈实验（如邀请同伴当面评价其社交表现），用真实外部信息直接挑战其核心不合理信念。

## Prompt

协同求助者共同选定1–2位可信、友善的同龄同伴（如同班/同宿舍同学）；在安全、私密且可控的咨询室内，当着求助者面，向同伴提出结构化问题（如‘她和你们说话时，是否真的像她说的那样脸红发抖、显得怪异？’）；引导求助者倾听、记录并即时复述他人反馈；随后与求助者一起对比其原有信念与实际反馈，标注差异，支持其表达认知松动（如‘原来没那么糟’）。

## Objective

修正核心自动思维
## Applicable Signals

- 求助者使用绝对化、以偏概全、灾难化语言描述他人反应（如‘他们一定会嘲笑我’）
- 求助者能识别该想法但无法自我反驳
- 求助者曾提及某位同伴态度友善或中立

## Contraindications

- 求助者当前极度羞耻或拒绝他人接触
- 缺乏可信同伴或环境不支持公开验证
- 信念内容涉及现实危险（如被伤害）而非认知扭曲

## Intervention Moves

- 共构验证问题（聚焦可观察行为，非主观评价）
- Third-party live testimony elicitation
- Real-time belief–evidence contrast mapping
- Cognitive softening reflection ('What felt different hearing that?')

## Workflow Steps

- Step 1: Collaboratively identify one concrete, observable belief (e.g., 'I look like a monster when speaking').
- Step 2: Select 1–2 trusted peers; obtain verbal consent for brief, respectful feedback.
- Step 3: Conduct in-session, face-to-face peer inquiry using behaviorally anchored questions (e.g., 'When she talks to you, do you notice her face turning red or hands shaking?').
- Step 4: Invite求助者 to paraphrase each response aloud; therapist mirrors and validates emotional impact.
- Step 5: Jointly note discrepancies between belief and observed evidence; elicit first-person cognitive shift statement.

## Constraints

- Must occur in-session with therapist present; never assigned as homework.
- Peer questions must target observable behaviors—not interpretations or feelings about the求助者.
- No more than two peers per session; total duration ≤12 minutes.

## Cautions

- Avoid leading questions or implying expectation of positive feedback.
- Monitor for shame escalation; pause and normalize if求助者 withdraws mid-process.
- Do not use peers who are authority figures or emotionally entangled with求助者.

## Output Contract

- 求助者 verbally articulates at least one objective peer observation (e.g., 'She said I didn’t shake or stutter'), and expresses emergent cognitive flexibility (e.g., 'Hmm… maybe it’s not as bad as I thought' or 'I didn’t realize they didn’t notice that').

## Example Therapist Responses

### Example 1

- Client/Input: 求助者坚信：‘我说话时他们肯定觉得我很蠢，会背地里笑话我。’
- Therapist/Output: 同伴回应：‘她说话挺清楚的，我们都没觉得奇怪，也没人笑话她。’ 求助者轻声说：‘哦……原来他们根本没注意这个。’
- Notes: Focus on observable absence of expected negative behavior.

### Example 2

- Client/Input: 求助者认为：‘我一开口，大家就会觉得我是个怪物。’
- Therapist/Output: 同伴回应：‘她就是有点慢热，但聊起来挺正常的，不像怪物。’ 求助者低头停顿后说：‘啊……在别人眼里，我还没有这么糟。’
- Notes: Direct quote from excerpt; marks early cognitive softening.

## Files

- `references/evidence.md`
- `references/evidence_manifest.json`

## Triggers

- 识别出具体、可验证的不合理信念（如‘他们会嘲笑我’）
- 求助者对信念真实性存疑但尚未动摇
- 有安全可控的同伴资源支持现场验证

## Examples

### Example 1

Input:

  求助者坚信：‘我说话时他们肯定觉得我很蠢，会背地里笑话我。’

Output:

  同伴回应：‘她说话挺清楚的，我们都没觉得奇怪，也没人笑话她。’ 求助者轻声说：‘哦……原来他们根本没注意这个。’

Notes:

  Focus on observable absence of expected negative behavior.

### Example 2

Input:

  求助者认为：‘我一开口，大家就会觉得我是个怪物。’

Output:

  同伴回应：‘她就是有点慢热，但聊起来挺正常的，不像怪物。’ 求助者低头停顿后说：‘啊……在别人眼里，我还没有这么糟。’

Notes:

  Direct quote from excerpt; marks early cognitive softening.