Skip to content
NewA calmer, faster ZeroHour

TraceVIC: Causal Reasoning over Code Evolution for Identifying Vulnerability-Inducing Commits

arXiv cs.AI / cs.LG / cs.CLby Fnu Tanish ()1 min read2
AI summary · grok-4.7Summaries by Vani AI

TraceVIC ranks vulnerability-inducing commits by reasoning over how vulnerable code evolves across revisions.

TraceVIC identifies vulnerability-inducing commits by localizing likely root-cause lines, tracing them through revision history, and ranking commits by their contribution to the vulnerable condition instead of picking the earliest or latest git-blame change. Temporal graph edges preserve correspondences between program elements across consecutive revisions. Ablations show modeling the full history improves F2 from 0.637 to 0.814. The authors report up to a 28.7% F2 gain over prior methods and a valid inducing commit for 78 of 79 vulnerabilities in four unseen C/C++ projects.

Key facts
  • Rejects git-blame earliest-or-latest heuristics for finding inducing commits.
  • Temporal graphs connect vulnerability-relevant code across revisions.
  • Full-history modeling raises F2 from 0.637 to 0.814.
  • A valid inducing commit was found for 78 of 79 cases.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

1 min read at arxiv.org

Software vulnerabilities are often discovered long after they are introduced, making it difficult to identify the vulnerability-inducing commit (VIC) responsible for introducing the underlying vulnerable condition. Existing VIC identification techniques largely rely on git blame to trace vulnerable code through revision history and use positional heuristics, such as selecting its earliest or most recent modification.

arXiv cs.AI / cs.LG / cs.CL

By Fnu Tanish · arxiv.org

Read the full article at arXiv cs.AI / cs.LG / cs.CL

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.