ontograph/ontoindex/ontoindex-guide

OntoIndex Reference

Read ontoindex-navigation first; then choose the exploring, debugging, impact-analysis, pr-review, refactoring, or cli recipe skill for the task. Pass repo: '<repo>' in every tool call.

Resources: ontoindex://repos; ontoindex://repo/{name}/context, /schema, /processes, /clusters, and /process/{processName} (append each suffix to the repository URI).

Nodes include File, Folder, Function, Class, Interface, Method, CodeElement, Community, Process, and language-specific types such as Struct, Enum, Property, and Constructor. Symbols expose name, filePath, startLine, and endLine; consult the schema resource for type-specific properties. All edges use the CodeRelation table with a type property (and confidence, reason, step). Edge types include CONTAINS, DEFINES, CALLS, IMPORTS, EXTENDS, IMPLEMENTS, HAS_METHOD, HAS_PROPERTY, ACCESSES, METHOD_OVERRIDES, METHOD_IMPLEMENTS, MEMBER_OF, and STEP_IN_PROCESS.

Use a bounded query (LIMIT must be at most 5000):

search({
  repo: '<repo>',
  action: 'cypher',
  query:
    "MATCH (caller)-[:CodeRelation {type: 'CALLS'}]->(f:Function {name: '<symbol>'}) RETURN caller.name, caller.filePath LIMIT 20",
});

Avoid variable-length Cypher patterns; use impact with maxDepth for traversal.