Version: 2.5.0 Status: ✅ Production Ready - All Core Features Implemented License: MIT

Multi‑platform clinical mental‑health LLM prompts and skills for screening, scoring, safety, intake, treatment planning, documentation, sleep diary, and opioid MME.

This is a clinical support tool, NOT a replacement for professional judgment:

• Does not diagnose mental health conditions
• Does not replace comprehensive clinical assessment
• Requires interpretation by licensed mental health professionals
• All interventions must be guided by clinical judgment and applicable laws/regulations
• Users are responsible for adhering to their scope of practice and professional standards

For crisis situations:

• 988 Suicide & Crisis Lifeline: Call or text 988
• Crisis Text Line: Text HOME to 741741
• Emergency Services: 911

The Clinical Toolkit provides structured frameworks for common clinical tasks in mental health practice:

• Screening & Assessment: Validated instruments for depression, anxiety, trauma, substance use, and suicide risk
• Clinical Workflows: Intake interview guides, treatment planning frameworks, documentation templates
• Evidence-Based: All instruments are validated, public domain tools with established psychometric properties
• Measurement-Based Care: Support for systematic outcome tracking and treatment adjustment

• Installation Guide
• Consumer Chatbot Guide
• Multi-Platform Testing Guide
• Manifest System

Developer / Agent Platforms

# Generate platform manifests for IDE/CLI agents
npm run generate:manifests

For platform-specific install commands and team setup, see:

• Installation Guide
• Manifest System

Consumer Chatbots (ChatGPT, Gemini, Perplexity, Claude Consumer, Ollama)

# Build all consumer chatbot bundles
npm run generate:consumers

# Validate outputs
npm run validate:consumers

# Optional: build Claude consumer zip files (one skill per zip)
npm run package:claudeai

Consumer bundles are written to dist/consumer/. For step-by-step setup, see:

• Consumer Chatbot Guide

📥 Download Latest Release

Download individual skill .zip files from the release assets. Claude.ai requires exactly one SKILL.md per upload.

Installation:

1. Download the skills you need (e.g., depression-screening.zip, anxiety-screening.zip)
2. Go to Claude.ai → Settings → Capabilities
3. Enable the Skills toggle
4. Click Upload skill and select a .zip file
5. Repeat for each skill you want to use

Available individual skills:

• depression-screening.zip - PHQ-9 and PHQ-2 assessments
• anxiety-screening.zip - GAD-7 and GAD-2 assessments
• trauma-screening.zip - PCL-5 and PC-PTSD-5 assessments
• substance-screening.zip - AUDIT-C and DAST-10 assessments
• suicide-screening.zip - ASQ and C-SSRS protocols
• intake-interview.zip - HEADSS and biopsychosocial templates
• treatment-planning.zip - SMART goals and ASAM/LOCUS criteria
• documentation.zip - SOAP and DAP note templates

Download clinical-toolkit-complete.zip for multi-skill plugin support:

# Download and install complete package
claude install-plugin clinical-toolkit-complete.zip

For maintainers and contributors, releases are managed through GitHub Actions:

• Manual trigger - On-demand releases when ready
• Automated packaging - All skills bundled into single ZIP
• Semantic versioning - Clear version numbering (1.0.0, 1.1.0, etc.)
• Auto-generated changelog - From commit history

For detailed release process, see Release Process Documentation.

Skills can be invoked in multiple ways:

1. Direct skill invocation:

/depression-screening

2. Natural language requests:

"Help me screen this patient for depression"
"I need to administer a PHQ-9"
"Can you guide me through a suicide risk assessment?"

3. Task-specific requests:

"Score this GAD-7 for me: [scores]"
"Help me write SMART goals for this patient"
"Create a SOAP note from this session"

1. Initial Screening:

   "I need to screen a new patient for depression and anxiety"
   → Uses depression-screening and anxiety-screening skills
   

2. Interactive Administration:

   "Let's start a PHQ-9 assessment"
   → Uses depression-screening skill in item-by-item mode
   

3. Risk Assessment:

   "The PHQ-9 Item 9 was positive - help me assess suicide risk"
   → Uses suicide-screening skill with C-SSRS or ASQ
   

4. Treatment Planning:

   "Based on these assessment results, help me create a treatment plan"
   → Uses treatment-planning skill for goals and level of care
   

5. Documentation:

   "Help me write a SOAP note for this intake session"
   → Uses documentation skill
   

Suicide Risk Screening:

• ANY positive response to suicidal ideation items requires immediate clinical follow-up
• Skills include escalation pathways and crisis resources
• Safety protocols embedded in relevant skills

Mandatory Reporting:

• Skills remind clinicians of mandatory reporting obligations
• Documentation guidance for reportable concerns

Scope of Practice:

• All skills emphasize the need for licensed professional judgment
• Tools support but do not replace clinical assessment

• Active suicidal ideation with plan and intent
• Homicidal ideation with identified target
• Acute psychosis with safety concerns
• Severe substance intoxication or withdrawal
• Child/elder abuse concerns

Always err on the side of safety and seek consultation when uncertain.

All instruments included are public domain or freely available:

• PHQ-9/PHQ-2 & GAD-7/GAD-2: https://www.phqscreeners.com
• PC-PTSD-5 & PCL-5: https://www.ptsd.va.gov
• C-SSRS: https://cssrs.columbia.edu
• ASQ: https://www.nimh.nih.gov/asq

Shared Clinical References (docs/references/):

• Crisis Protocols - Universal crisis response procedures
• Referral Guidelines - Cross-domain referral guidance
• Documentation Standards - Clinical documentation best practices
• Legal & Ethical Guidelines - HIPAA, informed consent, professional standards

Project Documentation:

• Clinical References - Scoring tables and interpretation guidelines

• American Psychological Association (APA)
• National Association of Social Workers (NASW)
• American Counseling Association (ACA)
• Substance Abuse and Mental Health Services Administration (SAMHSA)
• National Institute of Mental Health (NIMH)

clinical-toolkit/
├── .claude-plugin/
│   └── manifest.json              # Generated: Claude Code manifest
├── .github/copilot/
│   └── manifest.json              # Generated: GitHub Copilot manifest
├── .cursor/
│   └── manifest.json              # Generated: Cursor IDE manifest
├── .gemini/
│   └── manifest.json              # Generated: Gemini CLI manifest
├── .codex/
│   └── manifest.json              # Generated: OpenAI Codex manifest
├── manifest/
│   ├── manifest.json              # ⭐ Master manifest (edit this)
│   ├── templates/                 # Platform-specific templates
│   ├── overrides/                 # Platform-specific customizations
│   └── README.md                  # Manifest system documentation
├── scripts/
│   └── generate-manifests.js      # Manifest generation script
├── skills/                        # Skill implementations
│   ├── depression-screening/
│   │   ├── SKILL.md              # Quick reference + embedded essentials
│   │   ├── assets/               # Complete assessment files (PHQ-9, PHQ-2)
│   │   └── references/           # Topic-based guidance (severity, decision trees)
│   ├── anxiety-screening/         # (Same structure: assets + references)
│   ├── trauma-screening/          # (Same structure: assets + references)
│   ├── substance-screening/       # (Same structure: assets + references)
│   ├── suicide-screening/         # (Same structure: assets + references)
│   ├── intake-interview/
│   ├── treatment-planning/
│   └── documentation/
├── docs/
│   ├── references/                # Shared clinical references
│   │   ├── crisis-protocols.md
│   │   ├── referral-guidelines.md
│   │   ├── documentation-standards.md
│   │   └── legal-ethical-guidelines.md
│   └── clinical-references.md     # Shared reference materials
├── package.json                   # NPM scripts for manifest generation
└── README.md

Screening Skills (depression, anxiety, trauma, substance, suicide) use a modular structure:

• SKILL.md: Embedded essentials with quick reference tables, delegates to detailed files
• assets/: Complete assessment files (full items, scoring, documentation templates)
• references/: Topic-based references (severity levels, screening comparison, clinical decision trees)

Workflow Skills (intake-interview, treatment-planning, documentation) use single SKILL.md files for straightforward procedural guidance.

Each SKILL.md includes:

# [Skill Name]
## Description
## Quick Reference: Assessment Selection (screening skills)
## Usage
## Quick Reference: Severity Levels (screening skills)
## Assessment Administration
## Clinical Decision-Making
## Safety Protocols
## Documentation
## Limitations & Considerations
## Additional Resources (links to assets/ and references/)
## References

To add a new assessment to an existing skill:

1. Verify it's public domain or freely available
2. Add to the appropriate skill's "Available Assessments" table
3. Include full instrument items with proper attribution
4. Add scoring algorithm and interpretation guidelines
5. Include safety protocols if applicable
6. Cite validation studies and sources
7. Update clinical-references.md

1. Create skill directory: skills/[skill-name]/
2. Create SKILL.md using the standard template
3. Register in master manifest: Edit manifest/manifest.json:

   {
     "skills": [
       {
         "id": "skill-name",
         "path": "skills/skill-name/SKILL.md",
         "name": "Skill Display Name",
         "description": "Brief description including specific tool names",
         "category": "screening",
         "tags": ["tag1", "tag2"]
       }
     ]
   }

4. Generate platform manifests:

   npm run generate:manifests

5. Include both generic and specific terms in description for discoverability
6. Document in clinical-references.md if applicable
7. Commit all changes (including generated manifests)

We welcome contributions that:

• Add new validated, public domain assessment instruments
• Improve existing skill content with evidence-based information
• Enhance safety protocols and clinical guidance
• Fix errors or clarify instructions
• Add culturally-adapted versions of instruments

Before contributing:

• Ensure instruments are public domain or freely available
• Include proper citations and attributions
• Follow existing skill template structure
• Emphasize safety protocols where applicable
• Maintain non-judgmental, professional language
• Include evidence base (validation studies, clinical guidelines)

Not accepted:

• Proprietary or copyrighted instruments
• Unvalidated or experimental tools
• Content that could be harmful if misapplied
• Medical advice or prescriptive clinical decisions

This project uses an automated manifest generation system to support multiple AI platforms from a single source of truth.

Supported Platforms:

• ✅ Claude Code (.claude-plugin/)
• ✅ GitHub Copilot (.github/copilot/)
• ✅ Cursor IDE (.cursor/)
• ✅ Gemini CLI (.gemini/)
• ✅ OpenAI Codex (.codex/)

How it works:

1. Single source: Edit manifest/manifest.json only
2. Generate all: Run npm run generate:manifests
3. Automatic sync: All platform manifests are updated

Common commands:

# Generate all platform manifests
npm run generate:manifests

# Preview without writing files
npm run generate:manifests:dry-run

# Generate for specific platform
npm run generate:claude
npm run generate:copilot
npm run generate:cursor

Benefits:

• ✅ DRY principle - edit once, deploy everywhere
• ✅ Platform-specific overrides when needed (manifest/overrides/)
• ✅ Generated files committed for immediate use
• ✅ CI/CD validation ensures manifests stay in sync

For detailed documentation, see manifest/README.md

Before marking a skill as complete:

• All instrument items included and accurate
• Scoring algorithms verified against source materials
• Interpretation guidelines cite evidence base
• Safety protocols comprehensive and prominent
• Limitations clearly stated
• Citations complete and accurate
• Language professional and non-stigmatizing
• Manifests regenerated: npm run generate:manifests
• Multi-platform testing completed (see Multi-Platform Testing Guide)

• Project scaffolding complete
• Implement depression-screening (PHQ-9, PHQ-2)
• Implement anxiety-screening (GAD-7, GAD-2)
• Create shared docs/references/ infrastructure

• Implement trauma-screening (PC-PTSD-5, PCL-5)
• Implement substance-screening (AUDIT-C, DAST-10)
• Implement suicide-screening (C-SSRS, ASQ)
• Establish modular assets/references structure

• Implement intake-interview frameworks (HEADSS, Biopsychosocial)
• Implement treatment-planning tools (SMART goals, ASAM/LOCUS)
• Implement documentation templates (SOAP, DAP notes)

• Multi-platform manifest system (Claude, Copilot, Cursor, Gemini, Codex)
• Automated release workflow with GitHub Actions
• Complete packaging for Claude.ai distribution
• Comprehensive documentation

• Add culturally-adapted instrument versions
• Additional validated screening tools
• Integration with measurement-based care platforms
• Multi-language support

MIT License - See LICENSE file for details

All included instruments are public domain or freely available for clinical use:

• PHQ-9, PHQ-2, GAD-7, GAD-2: Public domain, no copyright restrictions
• PC-PTSD-5, PCL-5: Public domain (VA National Center for PTSD)
• C-SSRS: Free with attribution (Columbia University)
• ASQ: Public domain (NIMH)
• AUDIT-C: Freely available (WHO adaptation)
• DAST-10: Freely available for clinical use

Proper attribution is maintained for all instruments per their respective requirements.

Issues & Questions:

• GitHub Issues: github.com/rhavekost/clinical-toolkit/issues

Professional Consultation:

• This toolkit does not provide clinical consultation
• Seek supervision from licensed professionals for clinical questions
• Contact professional organizations for scope of practice guidance

This project builds on the foundational work of researchers and clinicians who developed and validated these instruments, making them freely available to improve mental health care.

Special recognition to:

• Drs. Robert L. Spitzer, Janet B.W. Williams, and Kurt Kroenke (PHQ/GAD instruments)
• VA National Center for PTSD (trauma screening instruments)
• Columbia Lighthouse Project (C-SSRS)
• National Institute of Mental Health (ASQ)
• All researchers who conducted validation studies

For questions about using these skills, see the individual skill documentation. For clinical emergencies, call 988 or 911.

Last Updated: 2026-02-03 Version 2.5.0 - Production Ready