lanternade-cad

List Modules

List all available CAD API modules. Returns module names, descriptions, and export counts. Use this first to discover what APIs are available.

Describe Module

Get detailed API documentation for a specific module. Returns type definitions, function signatures, and usage examples.

Search Api

Search for APIs by keyword. Returns matching functions, types, and classes with relevance scores. Use this to find specific functionality.

Execute Code

Execute TypeScript/JavaScript code against CAD APIs in a sandboxed environment. Use discovery tools first to understand available APIs.

List Entities

List CImux-DG entities (primitives, composers, investigators) with optional filtering. Token cost: ~100-500 depending on detail level. Use this for progressive disclosure: 1. First call with detailLevel: 'minimal' to get IDs 2. Then describe_entity for specific items Categories: - Primitives: file-system, syntax, pattern, relationship, metric, search - Composers: tier-0, tier-1, tier-2, tier-3, tier-4 - Investigators: structure-architecture, data-lifecycle, integration-operations, quality-requ

Describe Entity

Get detailed information about a specific CImux-DG entity. Token cost: ~500-1000. Returns full schema, examples, and related entities. Use after list_entities to drill down into specific items.

Search Entities

Search CImux-DG entities by keyword or semantic query. Token cost: ~200-500. Supports: - Keyword search: "dependency graph" - Semantic search: "find circular imports" - Combined: "security secret detection" Returns ranked results with relevance scores.

Run Primitive

Execute a single CImux-DG primitive. Token cost: varies by primitive output size. Common primitives: - P001: file-tree - enumerate files - P010: imports - extract import statements - P040: import-graph - build dependency graph - P050: complexity - cyclomatic complexity - P062: secret-scan - detect hardcoded secrets

Run Composer

Apply a CImux-DG composer to transform data. Token cost: varies by input size and composer type. Tiers: - T0 (Universal): merge, filter, map, reduce, union, intersect - T1 (Transform): simplify, sample, topN, truncate, normalize - T2 (Team): team-architecture, team-data, team-quality

Run Investigator

Execute a canonical investigator in a specific mode. Token cost: ~500-2000 depending on mode and scope. Priority investigators: - CI-01: Codebase Cartographer (structure mapping) - CI-02: System Architect (C4 diagrams) - CI-07: Data Architect (schema analysis) - CI-17: Validation Architect (error handling) - CI-23: Secret Scanner (security) Modes vary by investigator. Check describe_entity for available modes.

Build Pipeline

Build a reusable analysis pipeline from steps. Returns pipeline ID and resource estimates. Example pipeline definition: { "name": "architecture-audit", "steps": [ { "id": "s1", "type": "primitive", "entityId": "P001" }, { "id": "s2", "type": "primitive", "entityId": "P010", "composerChain": ["composer:simplify"] }, { "id": "s3", "type": "composer", "entityId": "composer:merge" } ], "scope": { "rootDir": "." } }

Validate Pipeline

Validate a pipeline before execution. Checks: - Entity availability - Dependency resolution - Estimated token cost

Execute Pipeline

Execute a built pipeline with optional checkpointing. Supports resumption from checkpoint for long-running pipelines.

Check Availability

Check if an entity can be executed. Returns: - canExecute: boolean - missingRequired: primitive IDs - aggregateFidelity: 0-5 quality score

Get Capability Stats

Get overall capability statistics. Returns counts and availability by category.