The world's first agentic reverse engineer.
Commits on main
fix: improve eval metrics with recall-weighted scoring and Ghidra type normalization
026770b Amruth committed 8 months ago
feat(cli): add kong eval command for scoring against ground truth
b09d005 Amruth committed 8 months ago
feat: add semantic synthesis module for cross-function analysis
b464b2c Amruth committed 8 months ago
feat: add syntactic normalizer for Ghidra decompilation cleanup
a4852ed Amruth committed 8 months ago
refactor: fix inline imports, move AnalysisStats to models, clean up init files
15c0050 Amruth committed 8 months ago
feat(cli): wire TUI as default mode, --headless for Rich output
31d9a3d Amruth committed 8 months ago
add type recovery by impl struct proposals and unification process
05ef03d Amruth committed 8 months ago
add/update tests for analyzer/triage agents and supervisor agent
e29da0d Amruth committed 8 months ago