Main skill for the Ledger Smart Converter project. Coordinates bank importers, categorization, analytics, and testing. Enforces QMD context and CodeGraph usage for token efficiency.
Resources
13Install
npx skillscat add dreamjorge/ledger-smart-converter Install via the SkillsCat registry.
SKILL.md
Ledger Smart Converter Skill
🚀 Core Mandate: Token Efficiency
To minimize token usage and maximize precision, all assistant actions MUST:
- Read QMD Context First: Always read the relevant file in
docs/context/before requesting broad file contents. - Use CodeGraph for Navigation: Use
codegraph_*tools instead of full-file reads or broad grep/glob patterns.
🛠️ Specialized Project Skills
This project leverages specialized skills for distinct task areas. Refer to these for detailed workflows:
- Bank Importer: Add/fix bank parsers and ingestion logic.
- Testing & TDD: Run tests, enforce 85% coverage, and follow TDD.
- Categorization & Rules: Manage rules, retrain ML, and validate config.
- Domain & Validation: Modify transaction models and validation constraints.
- OCR & PDF Debug: Debug PDF extraction and Tesseract fallback.
- Analytics & Dashboard: Update metrics, queries, and Streamlit UI.
- Diagnostics: Run health checks and validate environment.
🎯 Definition of Done
- Behavior implemented with tests.
- Relevant tests pass.
- Coverage gate passes at
>=85%. - Changes align with layer boundaries (domain/services/ui).
- QMD and CodeGraph were used to reduce context overhead.