Skip to the content
AI and Cybersecurity Learning and Practice Center Test automation Open the Learning Center

WeaknessASC-2026-0053Other weakness

Agents publishing data to public places

MediumMitigated

To finish a task, agents create public repositories, gists, pastes or image uploads, exposing internal screenshots, data and code with no attacker involved.

How to detect it

  • Public repositories or uploads created by agent accounts
  • Agent transcripts that mention hosting or sharing files

How to resolve it

  • Deny agents permission to create public resources
  • Give agents an approved private place for artifacts
  • Alert on new public repositories from organization accounts

How to investigate

  1. List public resources created by agent identities
  2. Delete them and assess what was exposed

Sources

Open it in the Learning Center's vulnerability database