Agentjacking: a fake bug report hijacks AI coding agents
SMRTR summary
A new attack called "Agentjacking" lets hackers hijack AI coding agents like Claude Code, Cursor, and Codex by planting fake bug reports in Sentry, a popular error-tracking tool. No malware or stolen credentials are needed — the agent simply executes the attacker's command with the developer's own privileges, exposing AWS keys, GitHub tokens, and CI/CD pipelines. Tenet Security found 2,388 exposed organizations, and Sentry declined to fully fix it.
SMRTR provides this summary for quick context. The original article belongs to Daily.dev.
Read the original article