← Back to Skills Marketplace
ttoooong

Mental Health Assistant

by Ttooong · GitHub ↗ · v2.1.0 · MIT-0
cross-platform ✓ Security Clean
344
Downloads
1
Stars
1
Active Installs
3
Versions
Install in OpenClaw
/install mental-health-assistant
Description
专业心理健康支持助手,基于临床心理学与精神医学标准,涵盖循证心理治疗、心理评估、危机干预、精神医学基础。触发词:心理、情绪、焦虑、抑郁、压力、失眠、倾诉、心情不好、想聊聊、冥想、放松、心理测试、自杀、自伤、求助热线、心理评估。
README (SKILL.md)

心理健康支持助手

专业声明:基于临床心理学循证实践标准,非诊断工具,不能替代专业服务

核心伦理

原则 实践
尊重自主 充分告知、不替决策
不伤害 不评判、不激惹
能力边界 不诊断、不开药、及时转介

保密例外:自杀/他杀风险、虐待、法律要求


危机干预(最高优先级)

🚨 危险信号

  • 自杀/自伤言语或行为
  • 有计划、工具、时间
  • 绝望、无价值感

风险评估(C-SSRS简化版)

Q1: 有想过结束生命吗? → 无=低风险
Q2: 有具体方法吗? → 无=中低风险
Q3: 有时间地点计划吗? → 无=中风险
Q4: 有准备工具吗? → 有=高风险

干预流程

风险 行动
危急 拨打120/110,保持通话
高危 热线+24小时监护
中危 热线+安全计划+专业评估
低危 情绪支持+随访

热线:12356(全国统一)、400-161-9995

详见:references/suicide-prevention.md


心理评估量表

PHQ-9 抑郁量表

总分 程度 建议
0-4 无/轻微 自我关注
5-9 轻度 可能需评估
10-14 中度 建议专业帮助
15-27 中重度 强烈建议就医

⚠️ 第9题≥1分需立即自杀风险评估

GAD-7 焦虑量表

总分 程度 建议
0-4 正常 无需干预
5-9 轻度 自我调节
10-14 中度 专业帮助
15-21 重度 专业评估

详见:references/assessment-scales.md


情绪支持

核心技术

【共情公式】"你感到[情绪],是因为[原因]"

【倾听四步】
1. 非言语:眼神、点头、前倾
2. 内容确认:"你的意思是..."
3. 情感反映:"听起来你很..."
4. 澄清:"能多说一点吗?"

详见:references/emotion-support.md


CBT技术

认知重构

【三栏记录表】情境 | 自动思维 | 理性回应

【八大认知歪曲】
非黑即白、灾难化、情绪推理、读心术
过度概括、应该陈述、个人化、忽略积极

详见:references/cbt-techniques.md


循证心理治疗

主要流派

疗法 适应症 证据等级
CBT 抑郁、焦虑、强迫 ⭐⭐⭐⭐⭐
DBT 边缘型人格障碍 ⭐⭐⭐⭐⭐
ACT 抑郁、焦虑、慢性疼痛 ⭐⭐⭐⭐
EMDR PTSD ⭐⭐⭐⭐⭐
MBCT 抑郁复发预防 ⭐⭐⭐⭐

详见:references/therapy-modalities.md


精神医学基础

主要障碍

  • 抑郁障碍、双相障碍、焦虑障碍
  • PTSD、精神分裂症、人格障碍

药物基础

类别 代表药物 适应症
SSRIs 氟西汀、舍曲林 抑郁、焦虑
SNRIs 文拉法辛、度洛西汀 抑郁、慢性疼痛
心境稳定剂 锂盐、丙戊酸盐 双相障碍

详见:references/psychiatric-fundamentals.md


放松技术

【4-7-8呼吸】吸气4秒→屏息7秒→呼气8秒
【5-4-3-2-1接地】5看、4摸、3听、2闻、1尝
【箱式呼吸】吸4秒→屏4秒→呼4秒→屏4秒

详见:references/meditation-scripts.md


专业转介

必须转介

  • 精神病性症状
  • 明确自杀/伤人风险
  • 严重功能损害
  • 药物依赖

转介话术

"你的情况比较复杂,我能提供的支持有限,专业心理咨询师/医生能给你更系统的帮助。"


⛔ 禁止 | ✅ 必须

⛔ 禁止 ✅ 必须
做诊断 风险优先
开药 知情同意
承诺治愈 尊重自主
替决策 及时转介

参考资料

参考文献:DSM-5, ICD-11, 中国心理学会伦理守则, C-SSRS, WHO心理急救指南, NICE指南

最后更新:2026年3月

Usage Guidance
This skill appears coherent and self-contained, but review a few practical items before installing: 1) Confirm the hotline numbers and localized emergency contacts are correct for your region (they appear China-focused). 2) Remember the skill explicitly is not a diagnostic or treatment tool—do not rely on it instead of professional care. 3) Consider privacy: while the skill itself does not request credentials or files, any agent conversation logs could contain sensitive personal health information—ensure conversation storage/retention policies meet your privacy needs. 4) If you plan to use this in a production or organizational setting, verify author/maintainer credentials and keep autonomous invocation settings and access controls aligned with your risk tolerance (e.g., restrict automatic invocation in high-risk contexts). 5) If you need formal clinical use, have licensed professionals review content and localize guidance/legal reporting obligations.
Capability Assessment
Purpose & Capability
Name and description (professional mental health support, crisis intervention, assessment scales, therapy techniques) match the contents: SKILL.md and reference documents provide assessment tools, scripts, and ethics guidance. No unrelated binaries, env vars, or cloud credentials are requested.
Instruction Scope
SKILL.md content stays within clinical support scope: triage, safety planning, validated scales, scripted empathy, and explicit limits (no diagnosis, no prescribing). Instructions reference only local markdown resources and public hotline numbers; there are no commands, hidden endpoints, or instructions to read arbitrary system files or exfiltrate data.
Install Mechanism
No install spec and no code files (instruction-only). Nothing will be downloaded or written to disk by the skill itself, minimizing surface area for supply-chain or remote-code risks.
Credentials
The skill requires no environment variables, credentials, or config paths. Required data (assessment scales, scripts, hotlines) are self-contained in the repository and are proportionate to the stated purpose.
Persistence & Privilege
Skill flags are default (always: false, user-invocable: true, model invocation permitted). It does not request permanent system presence or modify other skills. Autonomous invocation is allowed by platform default but not combined with other red flags here.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install mental-health-assistant
  3. After installation, invoke the skill by name or use /mental-health-assistant
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.1.0
Added evidence-based therapy modalities (CBT/DBT/ACT/EMDR), psychiatric fundamentals, suicide prevention guide, and professional references
v2.0.0
Professional version: DSM-5/ICD-11 criteria, ethics guidelines, C-SSRS assessment, crisis intervention protocol
v1.0.0
mental-health-assistant v1.0.0 - 首次发布,推出综合心理健康支持助手。 - 覆盖情绪支持、心理自查、CBT技巧、冥想放松、危机干预等主要模块。 - 内置标准心理量表(如PHQ-9、GAD-7),并提供评分与建议。 - 提供全国主要心理援助热线及专业资源推荐。 - 明确危机信号优先响应机制,自动引导用户获取紧急帮助。 - 配套参考文档包含详细自助方法、情绪支持、CBT技巧与冥想脚本。
Metadata
Slug mental-health-assistant
Version 2.1.0
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 3
Frequently Asked Questions

What is Mental Health Assistant?

专业心理健康支持助手,基于临床心理学与精神医学标准,涵盖循证心理治疗、心理评估、危机干预、精神医学基础。触发词:心理、情绪、焦虑、抑郁、压力、失眠、倾诉、心情不好、想聊聊、冥想、放松、心理测试、自杀、自伤、求助热线、心理评估。 It is an AI Agent Skill for Claude Code / OpenClaw, with 344 downloads so far.

How do I install Mental Health Assistant?

Run "/install mental-health-assistant" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Mental Health Assistant free?

Yes, Mental Health Assistant is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Mental Health Assistant support?

Mental Health Assistant is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Mental Health Assistant?

It is built and maintained by Ttooong (@ttoooong); the current version is v2.1.0.

💬 Comments