Security context

What an agent needs to avoid regressing past fixes and find the next vuln in this repo.

langchain-ai/langgraph

              main @ 8c9d59c
            

Fixes

CVEs

HIGH

Peak severity

57.1%

Coverage

Highlights

SQL Injection: 2 prior fixes. Scrutinize any change in this area.

Actions: most-fixed (1 issue). Treat as high-risk during review.

7 high-severity fixes in this history; regressions here are high-impact.

Recurring patterns

The bug types that recur here, drawn from past fixes, not open vulnerabilities.

SQL Injection: Unvalidated filter keys and limits in metadata queries allowed attackers to inject SQL commands into underlying SQLite queries during checkpointing operations.

Insecure Deserialization: Unpacking msgpack-encoded payloads without class restrictions allowed reconstructing arbitrary Python objects, leading to remote code execution. This is tracked under CVE-2026-28277.

Arbitrary Code Execution: Using complex deserialization utilities like jsonplus_serde on untrusted database metadata permitted executing arbitrary code instead of parsing pure JSON structures.