Connect your agent
AdvisoriesCriticalCheck your lockfilePluginHow it worksPricing Connect your agent

Advisory record · GHSA-pmch-g965-grmr

Langroid: SQLChatAgent _validate_query blocklist misses pg_read_file family enabling arbitrary file read

CVSS 8.7 HighEPSS 0.0069 · 53rd percentile

Published by the GitHub Advisory Database at 2 Jul 2026, 17:42 UTC. agentprime has not analysed this yet, so whether it affects any app is not known.

PackageEcosystemAffectedFixed in
langroidPyPI<0.64.00.64.0
Identifiers
GHSA-pmch-g965-grmr · CVE-2026-50180 · PYSEC-2026-2580
Weakness (CWE)
CWE-22 · CWE-89
Tags
Published
2 Jul 2026
Updated
13 Jul 2026

Machine-written summary

Summary written by Claude Haiku from the source advisory. Not reviewed by a person.

Langroid versions before 0.64.0 have a SQL validation bypass in SQLChatAgent: its regex blocklist omits PostgreSQL functions such as pg_read_file(), pg_stat_file() and pg_ls_logdir(), so a crafted SELECT can read arbitrary files or list directories on the PostgreSQL host. An attacker who can influence the LLM-generated SQL, such as through prompt injection in data the agent reads, can do this without database credentials, and the default configuration does not prevent it. Upgrade to langroid 0.64.0 or later.

Written 6 Nov 2026.

Not analysed yet

We haven’t published our analysis of this Advisory yet. Once a person has reviewed it, this page will say what’s vulnerable, what has to be true for your app to be affected, and how to check.

Severity and exploitation

ScoreRatingCVSSVectorSource
8.7 High 4.0 CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N CNA (GitHub_M)
  • CISA KEV Not listed.
  • EPSS 0.0069, the 53rd percentile, on 14 Nov 2026: FIRST’s estimate of the chance of exploitation in the next 30 days.

Timeline

  1. 31 Oct 2026, 15:01 UTC First seen by agentprime
  2. 31 Oct 2026, 15:01 UTC Confirmed by OSV

Sources

Facts on this page come from the GitHub Advisory Database (CC-BY 4.0), OSV and the other sources listed, as published. We don’t reproduce full advisory descriptions, which can contain exploit code.