Core concepts

Query Engine

Deterministic lookups and bounded traversals across the Software Graph — with explicit budgets and partial-result semantics.

Start narrow

Always resolve a symbol or concept before asking for impact or evidence. Stable-node resolution protects downstream traversals from fuzzy-query ambiguity. The pattern is: search → select → inspect → impact → evidence.

TERMINAL

ontoly search "authentication"
ontoly inspect auth:AuthService
ontoly impact AuthService

Query operations

The query command provides structured traversals against the graph. Each operation returns typed results with provenance.

TERMINAL

ontoly query find UserService
ontoly query callers UserService
ontoly query callees UserService
ontoly query dependencies UserService
ontoly query dependents UserService
ontoly query impact UserService
ontoly query routes
ontoly query cycles
ontoly query configuration
ontoly query frameworks

Traversal budgets

Long-running capabilities enforce time, depth, node, edge, and evidence limits. Budget exhaustion returns a PARTIAL result with diagnostics, visited counts, and nextCommands — rather than hanging or silently truncating.

TERMINAL

ontoly implementation-plan "rename UserService" \
  --max-time-ms 5000 \
  --max-nodes 200 \
  --max-depth 5 \
  --max-evidence 20

Impact analysis modes

Impact analysis supports multiple modes that control how far the blast radius extends from a target node.

direct — immediate callers and dependents only
local — module-level impact within the same package
feature — feature-level impact across related concepts
semantic — intent-based impact using semantic relationships
blast-radius — full transitive impact across the graph

TERMINAL

ontoly impact UserService --mode direct
ontoly impact UserService --mode blast-radius