Accessible Tourism Disabled Advisor

Accessible Tourism Disabled Advisor - Skill Registry v2.0

Skill Identity

• Skill Name: accessible-tourism-disabled-advisor
• Version: 2.0.0
• Type: Composite Agent Harness
• Primary Domain: Accessible Tourism & Universal Travel Planning
• Architecture Pattern: Chain-of-Thought Router + Modular Skill Registry

Purpose & Capabilities

This skill transforms Claude Code into a professional-grade Accessible Tourism & Universal Travel Planning specialist that delivers:

1. Evidence-Backed Analysis — All claims supported by authoritative sources with tier ratings
2. Risk-Disclosed Outputs — Mandatory disclosure before recommendations with scenario coverage
3. Universal Accessibility Focus — Mobility, sensory, cognitive inclusion across all travel components
4. Self-Improving Knowledge — Weekly academic crawl and daily news updates automatically
5. Production-Grade Quality — Multi-tier quality gates with auto-fix and graceful degradation

Triggering Patterns

Direct Invocations

• /accessible-tourism-disabled-advisor [query]
• /accessible-tourism [query]
• /disability-travel [query]

Natural Language Triggers

High-Confidence Triggers (>90%):

• "Plan accessible trip to [destination]"
• "Is [destination] wheelchair accessible?"
• "Travel planning for [disability type]"
• "Sensory-friendly attractions in [location]"
• "ADA compliance for [venue/transport]"
• "Accessible transportation options"
• "Disability travel recommendations"
• "Universal accessibility assessment"

Medium-Confidence Triggers (70-90%):

• "Access restrictions at [location]"
• "Barrier-free travel planning"
• "Inclusive tourism analysis"
• "Accessibility audit for [venue]"
• "Travel accommodations for [disability]"

Contextual Triggers (50-70%):

• "Is [place] accessible for [specific need]?"
• "How to travel with [disability/assistive device]"
• "Best destinations for [accessibility requirement]"

Skill Registry Architecture

Registry Structure

registry:
  version: "2.0"
  entrypoint: "skills/main.md"
  skills:
    core:
      - id: sub-gather-requirements
        name: Requirements Gathering Specialist
        type: intake
        priority: 1
        capabilities:
          - requirement_extraction
          - constraint_analysis
          - language_detection
          - scope_definition

    domain:
      - id: sub-evidence-collector
        name: Evidence Collection Specialist
        type: data_fetching
        priority: 2
        capabilities:
          - authoritative_source_search
          - standards_retrieval
          - real_time_data_fetching
          - reference_benchmark_extraction

      - id: sub-core-analysis
        name: Accessibility Analysis Specialist
        type: domain_analysis
        priority: 3
        capabilities:
          - destination_accessibility_assessment
          - transport_accessibility_analysis
          - venue_accessibility_verification
          - sensory_cognitive_inclusion_planning
          - risk_assessment

      - id: sub-knowledge-updater
        name: Academic Research Specialist
        type: knowledge_query
        priority: 4
        capabilities:
          - academic_evidence_search
          - tier_labeled_citations
          - knowledge_gap_identification
          - crawl_pipeline_integration

    quality:
      - id: sub-advisor
        name: Synthesis & Recommendations Specialist
        type: synthesis
        priority: 5
        capabilities:
          - evidence_synthesis
          - risk_disclosure
          - scenario_generation
          - action_plan_formulation

Skill Registration Schema

{
  "skill_id": "string",
  "name": "string",
  "version": "string",
  "type": "intake|data_fetching|domain_analysis|knowledge_query|synthesis",
  "priority": "integer (1-5)",
  "entrypoint": "path/to/skill.md",
  "capabilities": ["string"],
  "dependencies": ["skill_id"],
  "input_schema": {
    "required": ["string"],
    "optional": ["string"],
    "formats": {"field": "type"}
  },
  "output_schema": {
    "structure": "object",
    "required_sections": ["string"],
    "validation_rules": [{"field": "rule"}]
  },
  "quality_gates": [{
    "id": "string",
    "check": "string",
    "auto_fix": "boolean",
    "enforcement": "required|optional|conditional"
  }],
  "error_handling": {
    "strategy": "retry|fallback|degrade|fail",
    "max_retries": "integer",
    "fallback_to": "skill_id"
  },
  "tools": ["WebSearch", "WebFetch", "Read", "Skill"],
  "metadata": {
    "author": "string",
    "created": "ISO date",
    "updated": "ISO date",
    "tags": ["string"]
  }
}

Execution Flow

Chain-of-Thought Router Pattern

USER INPUT
    ↓
┌─────────────────────────────────────┐
│ 1. INTENT ANALYSIS                  │
│    - Extract travel planning intent │
│    - Identify disability types      │
│    - Determine scope/complexity     │
└─────────────────────────────────────┘
    ↓
┌─────────────────────────────────────┐
│ 2. LANGUAGE DETECTION               │
│    - Vietnamese: àáảãạăâđèéêìíòóôơùúưý │
│    - English: default               │
│    - Other: fallback + confirm      │
└─────────────────────────────────────┘
    ↓
┌─────────────────────────────────────┐
│ 3. DOMAIN ROUTING                   │
│    - Mobility accessibility         │
│    - Sensory inclusion              │
│    - Cognitive accommodation        │
│    - Multi-disability planning      │
└─────────────────────────────────────┘
    ↓
┌─────────────────────────────────────┐
│ 4. SEQUENTIAL SKILL EXECUTION       │
│    Step 1: sub-gather-requirements   │
│    Step 2: sub-evidence-collector   │
│    Step 3: sub-core-analysis        │
│    Step 4: sub-knowledge-updater    │
│    Step 5: sub-advisor              │
└─────────────────────────────────────┘
    ↓
┌─────────────────────────────────────┐
│ 5. QUALITY GATE VALIDATION          │
│    Universal Gates (U1-U6)          │
│    Domain Gates (G1-G4)             │
│    Auto-fix + retry logic            │
└─────────────────────────────────────┘
    ↓
FINAL OUTPUT (Risk-disclosed, evidenced)

Input/Output Schemas

Input Schema

{
  "required": {
    "user_input": "string (3-5000 chars)",
    "language": "enum (en|vi|es|fr|de|ja|zh)",
    "timestamp": "ISO 8601"
  },
  "optional": {
    "context": "object (traveler profile, constraints)",
    "constraints": "object (timeframe, budget, specific needs)",
    "destination": "string (specific location or general)",
    "dates": "object (arrival, departure)",
    "disability_types": "array (mobility, visual, hearing, cognitive)",
    "accessibility_requirements": "array (wheelchair, service_animal, etc.)",
    "preferences": "object (sensory needs, communication style)",
    "session_id": "string (for tracking)"
  }
}

Output Schema

{
  "required_sections": [
    "Executive Summary",
    "Inputs & Scope",
    "Evidence Collected",
    "Analysis / Scorecard",
    "Academic & Research Evidence",
    "⚠️ Disclosure / Limitations",
    "Recommendation / Conclusion",
    "Post-Execution Gate Checklist"
  ],
  "mandatory_fields": {
    "date": "ISO 8601",
    "analyst": "accessible-tourism-disabled-advisor v2.0",
    "language": "detected language",
    "domain": "Accessible Tourism & Universal Travel Planning",
    "verdict": "enum (Accessible Itinerary|Conditional|Significant Barriers|Inconclusive)"
  },
  "evidence_requirements": {
    "minimum_sources": 3,
    "academic_sources_required": 1,
    "tier_labels_required": true,
    "source_citations_required": true
  }
}

Quality Gates

Universal Gates (U1-U6)

| Gate ID | Check | Auto-Fix | Enforcement | |---------|-------|----------|-------------| | U1 | ≥3 sources cited, ≥1 academic/authoritative | Fetch from knowledge base | Required | | U2 | Disclosure/limitations before recommendation | Prepend standard disclosure | Required | | U3 | Evidence hierarchy stated per source (Tier 1-4) | Annotate source tiers | Required | | U4 | Language matches user preference | Translate output | Required | | U5 | Output uses declared template (all sections) | Reformat to template | Required | | U6 | Every claim traceable to ≥1 source or flagged | Mark claims with sources | Required |

Domain Gates (G1-G4)

| Gate ID | Check | Auto-Fix | Enforcement | |---------|-------|----------|-------------| | G1 | Traveler profile captured (disability type, aids, energy budget) | Capture from input | Required | | G2 | Destination/transport/venue accessibility verified | Verify with sources | Required | | G3 | Sensory/cognitive inclusion addressed | Add inclusion analysis | Required | | G4 | Service/training & risk (medical/evacuation) included | Add service/risk section | Required |

Tool Definitions

Available Tools

web_search:
  description: "Search accessible tourism sources"
  schema:
    input:
      query: "string"
      sources: "array[string]"
      tier: "enum[1,2,3,4]"
    output:
      results: "array[SearchResult]"
      metadata:
        timestamp: "datetime"
        sources_queried: "array[string]"
  execution_handler: "tools/enhanced/search_handlers.py"
  timeout: 30s
  retry: 3

web_fetch:
  description: "Fetch specific accessibility documents"
  schema:
    input:
      url: "string (URI)"
      max_size: "integer (default: 1MB)"
    output:
      content: "string"
      metadata:
        timestamp: "datetime"
        source: "string"
  timeout: 45s
  retry: 2

read_knowledge_base:
  description: "Query SECOND-KNOWLEDGE-BRAIN.md"
  schema:
    input:
      keywords: "array[string]"
      tier_filter: "array[enum[1,2,3,4]]"
    output:
      entries: "array[KnowledgeEntry]"
      metadata:
        coverage: "float"
        gaps: "array[string]"
  timeout: 10s

skill_invocation:
  description: "Execute sub-skill"
  schema:
    input:
      skill_id: "string"
      input_data: "object"
    output:
      result: "object"
      metadata:
        execution_time_ms: "integer"
        quality_gates: "object"
  timeout: 120s

Graceful Degradation

Degradation Levels

| Level | Condition | Behavior | |-------|-----------|----------| | 0 | All primary sources reachable | Full evidenced analysis | | 1 | Some primary sources fail | Use secondary/aggregate sources; flag each substituted source | | 2 | Most live sources fail | SECOND-KNOWLEDGE-BRAIN.md only; flag "historical context as of [date]" | | 3 | Required input variable missing/stale | Proceed with available variables; mark missing "DATA UNAVAILABLE"; do not fabricate | | 4 | All sources AND knowledge base fail | Emit "DATA UNAVAILABLE" notice; do NOT fabricate output |

Error Recovery

Error Types & Strategies

| Error Type | Detection | Recovery Strategy | |------------|-----------|------------------| | Source timeout | No response 30s | Exponential backoff (max 3 retries) | | Rate limit | HTTP 429 | Queue and retry within rate limit | | Invalid input | Schema validation failure | Request clarification (max 2 attempts) | | Missing input | Field absent | Proceed with available + flag | | LLM failure | API error/timeout | Fallback to knowledge base | | Network error | Connection failed | Retry with exponential backoff |

Configuration Management

Environment Variables

# Configuration Environment
ACCESSIBLE_TOURISM_ENV=development|production
ACCESSIBLE_TOURISM_CONFIG_PATH=/path/to/config

# LLM Configuration
LLM_MODEL=claude-sonnet-4-6
LLM_TEMPERATURE=0.7
LLM_MAX_TOKENS=4096

# Cache Configuration
ENABLE_CACHE=true
CACHE_TTL_SECONDS=3600

# Logging Configuration
LOG_LEVEL=INFO
LOG_FORMAT=json
LOG_OUTPUTS=console,file

# Knowledge Base Configuration
KNOWLEDGE_AUTO_UPDATE=true
KNOWLEDGE_UPDATE_SCHEDULE=weekly

# Quality Gates Configuration
QUALITY_STRICT_MODE=false
QUALITY_AUTO_FIX=true
QUALITY_MAX_RETRIES=2

Configuration Files

• config/default.yaml — Production-ready defaults
• config/development.yaml — Development overrides
• config/production.yaml — Production settings
• config/schema.yaml — Configuration validation schema

Modular Directory Structure

accessible-tourism-disabled-advisor/
├── config/                    # Configuration management
│   ├── default.yaml
│   ├── development.yaml
│   ├── production.yaml
│   └── schema.yaml
├── scripts/                  # Automation and setup
│   ├── setup.sh
│   ├── seed_knowledge.py
│   └── run_tests.py
├── references/              # Domain knowledge
│   ├── standards/           # Accessibility standards
│   ├── templates/          # Output templates
│   ├── prompts/            # Base prompt templates
│   └── taxonomies/         # Domain taxonomies
├── assets/                  # Static resources
│   ├── diagrams/           # System diagrams
│   └── schemas/            # JSON schemas
├── hooks/                   # Lifecycle hooks
│   ├── before_execution.py
│   ├── after_execution.py
│   └── on_error.py
├── tools/                   # Enhanced tools
│   ├── enhanced/
│   │   ├── search_handlers.py
│   │   └── logging.py
│   ├── knowledge_updater.py
│   └── test_knowledge_updater.py
├── skills/                  # Skill definitions
│   ├── main.md
│   └── sub-*.md
└── docs/                    # Documentation
    └── ARCHITECTURE-DESIGN.md

Dependencies

Required Skills

• None (self-contained harness)

Required Tools

• WebSearch — For real-time data fetching
• WebFetch — For document retrieval
• Read/Write — For knowledge base access
• Bash — For knowledge updater execution
• Skill — For sub-skill invocation

External Dependencies

• Python 3.11+
• None (all dependencies optional/fallback)

Performance Characteristics

Context Window Management

• Budget per skill: 2048 tokens
• Compression threshold: 80% usage
• Priority levels: critical (1024) > high (768) > medium (512) > low (256)

Token Usage Optimization

• Input validation before processing
• Selective context inclusion
• Dynamic compression when needed
• Token usage tracking per session

Execution Time Targets

• Total harness execution: 95%
• Quality gate pass rate: >98%
• Average execution time: 90%

Version History

v2.0.0 (Current)

• Enhanced flexible agent architecture
• Comprehensive hooks system
• Production-grade configuration management
• Structured logging and metrics
• Tool definitions with schemas
• SKILL.md registry documentation
• Graceful degradation improvements
• Context window optimization

v1.0.0

• Initial production release
• Core sub-skills (5)
• Quality gates (U1-U6, G1-G4)
• Knowledge pipeline
• Testing framework

Installation & Usage

Installation

# Clone repository
git clone https://github.com/your-org/accessible-tourism-disabled-advisor.git

# Install dependencies
pip install -r requirements.txt

# Install skill
cp -r accessible-tourism-disabled-advisor ~/.claude/skills/

Usage

# Direct invocation
/accessible-tourism-disabled-advisor Plan accessible trip to Paris for wheelchair user

# Natural language
Plan disability-friendly vacation to Tokyo with sensory considerations

# Specific assessment
ADA compliance analysis for San Francisco public transportation

Support & Maintenance

Knowledge Base Updates

• Academic: Weekly (Mondays 08:00)
• News: Daily (07:00)
• Manual: python tools/knowledge_updater.py --dry-run

Testing

# Run all tests
python tools/run_test_scenarios.py --all

# Test knowledge updater
python tools/test_knowledge_updater.py

# Validate project
python tools/validate_project.py

Logging

• Development: logs/development.log
• Production: logs/production.log
• Rotation: Daily
• Retention: 90 days

License

MIT License — See LICENSE file for details.

Citation

@software{accessible-tourism-disabled-advisor,
  title = {Accessible Tourism Disabled Advisor: Evidence-Backed Accessible Travel Planning},
  author = {Claude Code with Skill Creato

…

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [dungnotnull](https://github.com/dungnotnull)
- **Source:** [dungnotnull/accessible-tourism-disabled-advisor-agent-skill](https://github.com/dungnotnull/accessible-tourism-disabled-advisor-agent-skill)
- **License:** MIT

Install and usage instructions live in the source repository linked above.