What problem does it solve? When the ChromaDB vector index, the NetworkX knowledge graph, or their upstream inputs (codebase wiki, CDL files, curated YAML, parsers) are missing, there is no single procedure to rebuild the entire knowledge layer. This Skill orchestrates that full construction, covering both disaster-recovery reconstruction of an existing model's layer and bootstrapping a brand-new model into A2MC. ## Core Features & Use Cases - Two construction paths: Path R reconstructs an existing model's index from commit-pinned inputs (reproducibility contract); Path N bootstraps a new model (e.g., EcoSim, ReSOM) through 12 steps including per-model parsers, curated YAML, and loader registration. - Orchestrates sub-skills: Sequences generate-codebase-wiki, the build_rag_index.py invocation, inject-knowledge authoring discipline, and validate-rag-chain, while owning the glue none of them cover. - Verification gate: Proves both the vector index and the graph built correctly via get_stats and find_parameters_for_output checks, catching silent half-builds where the graph is empty. - Use Case: A collaborator wants to add EcoSim to A2MC. This Skill walks through creating the KB directory, generating the wiki, writing parameter/output parsers, authoring curated relationships YAML, building into a separate persist dir, and validating the chain. ## Quick Start Ask the agent to build the RAG and GraphRAG knowledge layer from scratch for your model, specifying whether you are reconstructing an existing model's index or onboarding a new model.
