Index multi-language codebases into hypergraphs for structural queries and HTML visualizations.
npx skills add https://github.com/denson/hypergraph_code_explorer --skill hce
Please help me install this Agent Skill. Skill: hce Source: https://github.com/denson/hypergraph_code_explorer/tree/main/skill Command: npx skills add https://github.com/denson/hypergraph_code_explorer --skill hce
Large or mixed-language codebases are hard to explore and reason about by reading files or feeding many sources to an LLM. HCE solves this by parsing code into a deterministic AST-backed hypergraph so structural questions (callers, inheritance, blast radius, subsystem boundaries) can be answered in milliseconds without costly token usage.
Install HCE, index your project's source root with skip-summaries, and run the visualize command to produce an interactive HTML visualization.
High-intent search queries and answers about installing and using this skill.
You can visualize code architecture by indexing a mixed-language repository into a hypergraph and generating a self-contained HTML visualization. This parses multiple languages into a single graph, enabling interactive structural exploration without reading source files.
Blast-radius impact analysis identifies all components affected by a code change by querying a prebuilt hypergraph index. It traces callers and dependencies across the codebase in milliseconds, avoiding the token cost of feeding many source files to an LLM.
Yes, multi-language indexing supports Python, JavaScript/TypeScript, Go, Rust, Java, C/C++, Ruby, and PHP. It parses mixed-language repositories into a single deterministic hypergraph index using tree-sitter extraction, with regex fallbacks for unsupported languages.
Generate interactive visualizations by accumulating evidence-backed tours for architecture explanations and exporting them as compact graph JSON. The pipeline produces self-contained HTML files with precomputed layouts, color palettes, and tour-driven spotlighting.
Parsing code into an AST-backed hypergraph is the best way to explore subsystem boundaries without high token usage. It answers structural queries like inheritance and caller lookups in milliseconds directly from a cached index, bypassing costly LLM context windows.
Yes, the indexing pipeline includes regex fallbacks for languages outside its native tree-sitter support. This ensures unsupported file types are still incorporated into the hypergraph index for broader dependency tracing and structural queries.
hafley66Explain tree-sitter's parsing model and node classification for CST vs AST debugging.
stribusTransform JSON modules and dependencies into interactive 3D architecture visualizations.
bonjohenGenerate architecture and evidence-graph diagrams from a project knowledge graph.
rohanbhatia27Construct and query multi-language source code graphs for security analysis.
