Claude Skills by tclatos

Reference for the Document Graph navigation tools (get_folder_toc, get_document_toc, get_section_content, search_sections, list_documents) and the Ladybug schema they traverse. Use to look up exact tool arguments, return shapes, and the Folder/Document/MarkdownSection node model when navigating a document corpus.

Answer questions over a Document Graph (Folders → Documents → Markdown sections) stored in a Ladybug database by navigating the heading hierarchy and reading only the relevant sections — hybrid agentic RAG (vector + keyword search). Use whenever the user asks about the content of ingested documents and you have the get_folder_toc, get_document_toc, get_section_content, search_sections, and list_documents tools available.

Add or modify the genai-graph CLI command groups — cli kg (create/info/schema/agent/cypher/query/view), cli docgraph (run/build/list/toc/cat/search/tui), and cli neo4j (analyze/convert/subset/import/query/info). Use when editing genai_graph/core/commands_*.py or genai_graph/neo4j_import/commands.py, adding a KG CLI subcommand, or wiring a command to the workflow engine.

Build and run the genai-graph Document Graph deep agent — genai_graph/agent/docgraph_agent.py (create_docgraph_agent, prepare_docgraph_agent, run_docgraph_agent, resolve_db_path, build_docgraph_system_prompt), the colocated runtime skills under genai_graph/agent/skills, the cli docgraph agent command, the deepseek_v4flash@openrouter LLM registry entry, and the docgraph profile. Use when editing the agent module, its runtime skills, the agent CLI, or wiring a downstream project (e.g. rfq_prici...

Build and navigate the genai-graph Document Graph — the Folder/Document/MarkdownSection provenance hierarchy, DocumentGraphFactory, tree_parser, ingest_document_graph/drop_document_graph, and the vectorless agentic RAG navigation tools. Use when editing genai_graph/kg/document_graph, writing a document-ingestion pipeline, or giving an agent tools to walk a document corpus by heading.

Work on the genai-graph Streamlit KG Explorer and webapp — the interactive Cypher query UI, Text-to-Cypher (natural language) panel, D3 graph visualization, tabular results, page assembly, and shared UI components. Use when editing genai_graph/webapp or genai_graph/main/streamlit.py, adding a KG Explorer page, or debugging the Streamlit KG UI.

Export genai-graph Knowledge Graph artifacts — interactive HTML/D3 schema diagrams and graph visualizations (generate_html, export_schema_html), DAG HTML, the consolidated Markdown warnings report (export_warnings), info/schema/JSON/parquet artifacts, and ResolvedSchema rendering. Use when editing genai_graph/kg/export, adding a viz/info/warnings output, or debugging a cli kg create export step.

Author genai-graph factories that turn a data source into graph nodes — JsonFileBackedFactory, TableBackedFactory, DocumentGraphFactory, MarkdownBamlFactory (inline BAML entity extraction), DocumentDirectoryFactory, Neo4jImportFactory, SimilarityFactory, canonical node reuse, and GraphRegistry multi-source merge. Use when adding a new data source to a Knowledge Graph, writing a factory subclass, or editing genai_graph/kg/factories.

Ingest data into a genai-graph Ladybug (Kuzu-compatible) backend — KuzuBackend, create_graph/restart_database, MERGE/upsert semantics, parquet cache + fingerprint invalidation, embeddings via EmbeddingsHandler, and HNSW vector-index drop/recreate around merge. Use when editing genai_graph/kg/ingest or genai_graph/kg/backend.py, debugging a kg create ingest failure, or tuning cache/force-stage behavior.

Import Neo4j JSONL exports into a genai-graph Ladybug database — SchemaAnalyzer, Neo4jToKuzuConverter, the cli neo4j analyze/convert/subset/import/query/info commands, and the typed Neo4jImportFactory with Neo4jNodeMapping/Neo4jRelationMapping. Use when migrating a Neo4j graph, editing genai_graph/neo4j_import, or wiring a Neo4j source into a multi-factory KG.

Query a genai-graph Knowledge Graph and build Cypher-aware agents — direct Cypher via the backend, text2cypher_chain/query_kg, the kg_cypher_query LangChain tool and build_kg_agent_system_prompt, CALL QUERY_VECTOR_INDEX semantic search with $query_vector, Document Graph navigation tools, and SimilarityFactory similarity relations. Use when editing genai_graph/kg/query, writing a KG agent, or wiring graph tools into an agent profile.

Navigate the genai-graph repository by mapping work areas to docs, code, config, and the matching kg-* skill. Use before making cross-cutting changes to genai_graph/ or when deciding which genai-graph domain skill to load. Load this first whenever a task touches Knowledge Graph construction, the Document Graph, Cypher agents, Neo4j import, KG workflows, or the KG Explorer.

Library-level procedures for maintaining a genai-graph schema across factories — assess whether a Pydantic field is dead code, merge/unify two node types across factories, add a node type or field, and diagnose schema warnings ("Class X referenced but has no GraphNode", label collisions, orphaned nodes, duplicate relationships). Use when tidying a multi-factory KG, interpreting a warnings report, or generalizing an Agents_Skills.md procedure past a single project.

Define and compile a genai-graph GraphSchema — GraphNode/GraphRelation, name_from/key_from/table_name, AUTO_ID, field-path auto-deduction, p_ edge-property exclusion, coherence warnings, ResolvedSchema rendering, and embedding index_fields. Use when authoring or debugging a Knowledge Graph schema (genai_graph/kg/schema) or when a kg create / factory build emits schema warnings.

Orchestrate genai-graph Knowledge Graph pipelines with the genai-tk workflow engine — kg_create_step/kg_build_step/docgraph_build_step, force stages (unzip/pdf/md/parquet/graph/embed/all), multi-factory pipelines, cli kg create vs cli docgraph run, and sub-workflow composition. Use when editing genai_graph/orchestration or config/workflows/*.yaml, or debugging a KG build pipeline.