Sustainable Archaeology Tourism

Sustainable Archaeology Tourism Planning Advisor

Skill Overview

This is a production-grade harness for Sustainable Cultural-Heritage Tourism Planning analysis. It transforms Claude into a domain expert that delivers structured, evidence-backed outputs through a sophisticated multi-agent architecture with dynamic skill routing, comprehensive error handling, and continuous knowledge improvement.

Architecture Version 2.0

Enhanced Architecture Features

• Dynamic Skill Routing: Context-aware chain-of-thought router for optimal skill selection
• Parallel Execution: Independent skills execute concurrently when possible
• Hook System: Lifecycle hooks for pre/post skill execution and error handling
• State Management: Distributed state with context propagation
• Schema Registry: JSON schemas for all skill inputs/outputs
• Graceful Degradation: 5-level fallback system with explicit limitation flags
• Knowledge Pipeline: Continuous crawl pipeline with deduplication and scoring
• Structured Logging: Comprehensive logging with severity levels and correlation IDs

Skill Registry

Core Skills

| Skill ID | Description | Input Schema | Output Schema | Parallel | Dependencies | |----------|-------------|--------------|---------------|----------|--------------| | main-harness | Primary orchestrator with quality gates | UserQuery | AnalysisReport | false | [sub-gather-requirements] | | sub-gather-requirements | Intake specialist for requirement clarification | RawUserMessage | Requirements | false | [] | | sub-evidence-collector | Data librarian for authoritative sources | Requirements | EvidenceBundle | false | [sub-gather-requirements] | | sub-core-analysis | Heritage tourism planner for capacity analysis | SiteEvidence | CoreAnalysis | false | [sub-evidence-collector] | | sub-knowledge-updater | Research librarian for academic evidence | AnalysisKeywords | KnowledgeEvidence | true | [] | | sub-advisor | Principal advisor for synthesis and recommendations | AnalysisInputs | RecommendationReport | false | [sub-core-analysis, sub-knowledge-updater] |

Skill Execution Modes

• Sequential: Strict order (main-harness, sub-gather-requirements)
• Parallel: Concurrent execution (sub-knowledge-updater can run with sub-core-analysis)
• Conditional: Execution based on context (adaptive routing based on analysis complexity)

Input/Output Schemas

UserQuery Schema

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "rawMessage": {"type": "string"},
    "providedContext": {"type": "object"},
    "language": {"type": "string", "enum": ["en", "vi", "auto"]},
    "timestamp": {"type": "string", "format": "date-time"}
  },
  "required": ["rawMessage"]
}

Requirements Schema

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "objectOfAnalysis": {"type": "string"},
    "scope": {"type": "array", "items": {"type": "string"}},
    "timeframe": {"type": "string"},
    "availableInputs": {"type": "array", "items": {"type": "string"}},
    "targetAudience": {"type": "array", "items": {"type": "string"}},
    "language": {"type": "string", "enum": ["en", "vi"]},
    "analysisType": {"type": "string", "enum": ["screening", "comprehensive", "focused", "comparative"]},
    "confirmedAt": {"type": "string", "format": "date-time"}
  },
  "required": ["objectOfAnalysis", "scope", "language", "analysisType"]
}

EvidenceBundle Schema

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "currentData": {"type": "array"},
    "authoritativeDocs": {"type": "array"},
    "recentDevelopments": {"type": "array"},
    "referenceBenchmarks": {"type": "array"},
    "evidenceQuality": {"type": "string"},
    "collectedAt": {"type": "string", "format": "date-time"},
    "limitationFlags": {"type": "array"}
  },
  "required": ["currentData", "authoritativeDocs"]
}

AnalysisReport Schema

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "reportId": {"type": "string"},
    "executiveSummary": {"type": "string"},
    "inputsAndScope": {"type": "object"},
    "evidenceCollected": {"type": "object"},
    "analysisScorecard": {"type": "object"},
    "actionPlan": {"type": "array"},
    "academicEvidence": {"type": "array"},
    "disclosure": {"type": "string"},
    "conclusion": {"type": "string", "enum": ["Sustainable Plan", "Conditional (caps)", "Over-Tourism Risk", "Inconclusive"]},
    "scenarios": {"type": "array"},
    "keyRisks": {"type": "array"},
    "evidenceChain": {"type": "array"},
    "remediation": {"type": "array"},
    "gatesPassed": {"type": "array"},
    "degradationLevel": {"type": "integer", "minimum": 0, "maximum": 4},
    "generatedAt": {"type": "string", "format": "date-time"}
  },
  "required": ["executiveSummary", "conclusion", "disclosure", "gatesPassed"]
}

Hook System

Lifecycle Hooks

| Hook Name | Trigger | Purpose | Example Use | |-----------|---------|---------|-------------| | pre-skill-execution | Before any skill executes | Logging, validation, context enrichment | Log execution start, validate inputs | | post-skill-execution | After any skill completes | Result processing, state updates | Cache results, update state | | on-skill-error | When a skill fails | Error recovery, fallback activation | Activate degradation level | | pre-quality-gate | Before quality gate evaluation | Gate preparation, data gathering | Assemble gate check data | | post-quality-gate | After quality gate evaluation | Result processing, auto-fix | Apply auto-fix procedures | | on-report-generation | Before final report output | Formatting, template application | Apply output template | | post-report-generation | After final report output | Cleanup, logging | Log completion, cleanup resources |

Hook Registration

Hooks are registered in config/hooks.json with the following schema:

{
  "hookName": {
    "enabled": true,
    "priority": 1,
    "handler": "path/to/handler/script.py",
    "config": {}
  }
}

Quality Gates

Universal Gates (U1-U6)

• U1: ≥3 sources cited, ≥1 academic/authoritative
• U2: Disclosure/limitations before recommendation
• U3: Evidence hierarchy stated per source (Tier 1–4)
• U4: Language matches user preference
• U5: Output uses declared template (all sections)
• U6: Every claim traceable to ≥1 source or flagged

Domain Gates (G1-G4)

• G1: Carrying capacity determined
• G2: Conservation/visitation balance (zones/caps)
• G3: Impact monitoring present
• G4: Community & interpretation addressed

Gate Enforcement

Gates are evaluated with auto-fix logic:

1. Apply gate check
2. On failure, run auto-fix procedure
3. Retry gate check (max 2 retries)
4. If still failing, emit explicit limitation notice
5. Continue to next gate (fail-open for safety)

Error Handling & Graceful Degradation

Degradation Levels

| Level | Condition | Behavior | Limitation Notice | |-------|-----------|----------|-------------------| | 0 | All primary sources reachable | Full evidenced analysis | None | | 1 | Some primary sources fail | Use secondary/aggregate sources; flag substituted sources | Level 1 banner | | 2 | Most live sources fail | SECOND-KNOWLEDGE-BRAIN.md only; flag historical context | Level 2 banner | | 3 | Required input variable missing/stale | Proceed with available; mark missing "DATA UNAVAILABLE" | Level 3 banner | | 4 | All sources AND knowledge base fail | Emit "DATA UNAVAILABLE" notice; do NOT fabricate | Level 4 banner |

Error Recovery Matrix

| Error Type | Detection | Recovery | Retry Limit | Fallback | |-------------|-----------|----------|-------------|----------| | Source timeout | no response 30s | retry alternate source | 3 | Knowledge base | | Invalid input | out-of-range/schema mismatch | ask user to confirm | 2 | Proceed with flag | | Missing input | field absent | proceed with available + flag | N/A | Flag explicitly | | Stale reading | timestamp old | flag, request refresh | 1 | Use with flag | | Knowledge base miss | no matches | WebSearch gap-fill + queue crawl | 2 | WebSearch only | | Conflicting actions | mutually exclusive actions | apply precedence | N/A | Precedence rules | | Envelope unavailable | no setpoint for object/stage | use genus/category fallback | 1 | Fallback + flag | | Object/class ambiguous | classification unclear | ask user to confirm | 2 | Request clarification |

Context Window Management

Token Optimization Strategy

1. Progressive Loading: Load only essential content initially
2. Selective Reading: Use targeted Read with offset/limit for large files
3. Reference Caching: Cache commonly accessed references
4. Context Compression: Summarize intermediate outputs
5. Memory Management: Release unused context between skills

Token Budget Allocation

| Component | Budget | Priority | |-----------|--------|----------| | User query + requirements | 5% | Critical | | Evidence collection | 25% | High | | Core analysis | 30% | Critical | | Knowledge base query | 15% | Medium | | Synthesis & recommendation | 20% | Critical | | Quality gate overhead | 5% | Low |

State Management

Distributed State Pattern

State is maintained across skill executions with the following structure:

{
  "sessionId": "uuid",
  "correlationId": "uuid",
  "state": {
    "requirements": null,
    "evidenceBundle": null,
    "coreAnalysis": null,
    "knowledgeEvidence": null,
    "recommendation": null
  },
  "context": {
    "language": "auto",
    "degradationLevel": 0,
    "gatesPassed": [],
    "gatesFailed": [],
    "executionLog": []
  },
  "metadata": {
    "startedAt": "timestamp",
    "lastUpdatedAt": "timestamp",
    "currentSkill": "skill-id",
    "skillExecutionOrder": []
  }
}

Logging System

Log Levels

• TRACE: Detailed execution tracing
• DEBUG: Diagnostic information
• INFO: General informational messages
• WARN: Warning conditions
• ERROR: Error conditions
• CRITICAL: Critical conditions requiring immediate attention

Log Format

{
  "timestamp": "ISO8601",
  "level": "INFO",
  "correlationId": "uuid",
  "skill": "skill-id",
  "message": "Log message",
  "context": {},
  "error": {}
}

Log Outputs

• Console: Structured JSON logs to stdout
• File: Append to logs/skill-execution-{date}.log
• Error File: Separate error logs to logs/skill-errors-{date}.log

Knowledge Pipeline

Crawl Schedule

# Weekly academic update (Mondays 8:00 AM)
0 8 * * 1 python D:/972026/245-sustainable-archaeology-tourism/scripts/knowledge_updater.py >> logs/knowledge_update.log 2>&1

# Daily news update (Daily 7:00 AM)
0 7 * * * python D:/972026/245-sustainable-archaeology-tourism/scripts/knowledge_updater.py --news-only >> logs/knowledge_news.log 2>&1

Knowledge Sources

Tier 1 (Authoritative):

• UNESCO World Heritage Centre
• ICOMOS International Cultural Tourism Charter
• UNWTO Sustainable Tourism Criteria
• National heritage agencies

Tier 2 (Academic):

• Journal of Sustainable Tourism
• Tourism Management
• Journal of Cultural Heritage
• Annals of Tourism Research

Tier 3 (Industry):

• Industry reports
• Professional associations
• Technical guidelines

Tier 4 (News/Blog):

• Industry news
• Conference proceedings
• Working papers

Directory Structure

D:\972026\245-sustainable-archaeology-tourism\
├── SKILL.md                    # This file - comprehensive skill registry
├── CLAUDE.md                    # Project identity and instructions
├── PROJECT-detail.md            # Full technical specification
├── PROJECT-DEVELOPMENT-PHASE-TRACKING.md  # Build roadmap
├── README.md                    # Public documentation
├── SECOND-KNOWLEDGE-BRAIN.md    # Living knowledge base
├── skills/                      # Skill definitions
│   ├── main.md                  # Main orchestrator
│   ├── sub-gather-requirements.md
│   ├── sub-evidence-collector.md
│   ├── sub-core-analysis.md
│   ├── sub-knowledge-updater.md
│   └── sub-advisor.md
├── scripts/                     # Executable scripts
│   ├── knowledge_updater.py    # Knowledge crawl pipeline
│   ├── hook_handlers.py         # Hook implementation
│   ├── state_manager.py         # State management
│   ├── logger.py               # Structured logging
│   └── skill_router.py         # Dynamic skill routing
├── references/                  # Reference documentation
│   ├── hooks.md                # Hook system documentation
│   ├── schemas.md              # Schema definitions
│   ├── quality-gates.md        # Quality gate documentation
│   └── error-handling.md       # Error handling procedures
├── assets/                      # Static resources
│   ├── templates/               # Output templates
│   │   ├── report-template-en.md
│   │   └── report-template-vi.md
│   └── configs/                # Configuration schemas
│       ├── hooks-schema.json
│       └── logging-schema.json
├── config/                      # Configuration files
│   ├── hooks.json              # Hook registration
│   ├── skills.json             # Skill configuration
│   └── logging.json            # Logging configuration
├── tools/                       # Legacy tools (maintained)
├── tests/                       # Test scenarios and results
└── logs/                        # Log files

Usage

Basic Invocation

/sustainable-archaeology-tourism [user query]

Example Queries

• "Analyze sustainable tourism options for Angkor Archaeological Park"
• "Assess carrying capacity for My Son heritage site"
• "Evaluate visitor management strategies for Ha Long Bay"
• "Develop community-based tourism plan for Cu Chi tunnels"

Version History

• 2.0.0 (2026-07-27): Major architecture upgrade with dynamic routing, hook system, and enhanced error handling
• 1.0.0 (2026-07-13): Initial production release

License

MIT License - See LICENSE file for details

Contributing

This skill is part of the 972026-skill-library. Contributions should follow the library standard defined in D:\972026\SKILL-STANDARD.md.

Source & license

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

• Author: dungnotnull
• Source: dungnotnull/sustainable-archaeology-tourism-agent-skill
• License: MIT

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