Shan Snow Tea Production

SKILL.md — Skill Registry & Architecture

Skill Identity

Name: shan-snow-tea-production Tagline: Shan Tuyet (Ancient Tree) Tea Production Optimization — Evidence-backed analysis & decision-support harness Version: 1.0.0 Domain: Ancient-Tree Tea Processing & Quality Status: Production Ready

---

Skill Registration

Registration Protocol

Skills in this harness follow a strict registration protocol defined in the library standard (D:\972026\SKILL-STANDARD.md). Each skill must:

1. Declare Frontmatter: YAML frontmatter with name and description fields
2. Implement Required Sections: Role, Workflow, Tools, Output Format, Quality Gates
3. Follow File Naming: sub-{skill-name}.md for sub-skills, main.md for harness
4. Use Standard Output Templates: All outputs follow domain-specific templates
5. Implement Quality Gates: Universal (U1-U6) + Domain (G1-G4) gates with auto-fix

Skill Resolution

When a skill is invoked via Skill("skill-name"):

1. Path Resolution: Check skills/{skill-name}.md → fallback to {skill-name}/SKILL.md
2. Frontmatter Validation: Ensure name and description present and valid
3. Content Loading: Load skill content into context ( Dict[str, Any]:

""" Hook execution function.

Args: context: Current execution context with inputs and state config: Application configuration

Returns: Updated context or modified configuration """ # Hook logic here return context

---

## Configuration Management

### Environment Variables

| Variable | Purpose | Default |
|----------|---------|---------|
| `APP_ENV` | Environment (development/staging/production) | `production` |
| `DEBUG` | Enable debug mode | `false` |
| `LOG_LEVEL` | Logging level | `INFO` |
| `MAX_RETRIES` | API retry limit | `3` |
| `TIMEOUT` | API timeout (seconds) | `30` |

### Configuration Loading

```python
from config.settings import get_config, export_config_dict

# Get configuration
config = get_config()

# Access sub-configurations
domain_keywords = config.domain.keywords
api_timeout = config.api.default_timeout
quality_gates = config.quality.min_sources_required

# Export for inspection
config_dict = export_config_dict(config)

---

Knowledge Pipeline

Automated Updates

The knowledge base (SECOND-KNOWLEDGE-BRAIN.md) is updated automatically:

• Weekly academic update: Mondays 08:00
• Daily news update: Daily 07:00

Manual Execution

# Standard update
python tools/knowledge_updater.py

# Dry run (preview changes)
python tools/knowledge_updater.py --dry-run

# News only update
python tools/knowledge_updater.py --news-only

# Custom keywords
python tools/knowledge_updater.py --keywords "tea fermentation" "catechins"

Knowledge Structure

# SECOND-KNOWLEDGE-BRAIN.md

## 1. Core Concepts & Frameworks
## 2. Key Research Papers & Standards
## 3. State-of-the-Art Methods & Tools
## 4. Authoritative Data Sources
## 5. Analytical Frameworks
## 6. Self-Update Protocol
## 7. Knowledge Update Log

---

Testing & Validation

Test Scenarios

Located in tests/test-scenarios.md with 5+ production-grade scenarios:

1. Standard Analysis — Full inputs, all sources available
2. Minimal Input — Sparse inputs, test requirements gathering
3. Comparison Mode — Compare multiple processing approaches
4. Risk/Conflict — Conflicting recommendations, test disclosure
5. Degraded Mode — Limited sources, test graceful degradation

Running Tests

# Run all test scenarios
python tools/run_test_scenarios.py --all

# Run specific scenario
python tools/run_test_scenarios.py --scenario standard

# Validate project structure
python tools/validate_project.py

# Test knowledge updater
python tools/test_knowledge_updater.py

---

Performance Considerations

Context Window Management

• Target usage: 80% of context window
• Safety margin: 1000 tokens minimum
• Sub-skill content: <500 lines per skill file
• Knowledge base: Append-only, new entries first

Token Optimization

• Compact output templates: Use structured formats
• Selective tool use: Invoke tools only when needed
• Caching: Cache repeated fetches within session
• Lazy loading: Load references only when accessed

Rate Limiting

• API calls: 60 requests/minute
• Concurrent requests: Maximum 5
• Respect 429: Always handle rate limit responses
• Exponential backoff: Base delay 2s, max 60s

---

Integration Points

External Systems

• FAO: Tea production references and statistics
• ISO 3720: Tea quality standards
• Vietnam Tea Association: Local guidelines and practices
• Academic databases: ArXiv, Semantic Scholar, food chemistry journals

Internal Systems

• config/settings.py: Configuration management
• tools/knowledge_updater.py: Knowledge pipeline
• tools/runtestscenarios.py: Test orchestrator
• hooks/: Lifecycle management

---

Version History

| Version | Date | Changes | |---------|------|---------| | 1.0.0 | 2026-07-27 | Production-grade release with full architecture | | 0.x | Earlier | Development iterations |

---

References

• D:\972026\SKILL-STANDARD.md — Library-wide skill standard
• PROJECT-detail.md — Full technical specification
• PROJECT-DEVELOPMENT-PHASE-TRACKING.md — Build roadmap
• config/settings.py — Configuration implementation
• Reference implementation: D:\vn-finance-analysis-hd-skill

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/shan-snow-tea-production-agent-skill
• License: MIT

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