CVE-2026-119050: Agno PandasTools unsafe reflection vulnerability

Agno through version 3.1.11 contains an unsafe reflection vulnerability in PandasTools that allows attackers to invoke arbitrary pandas functions through prompt injection, enabling code execution or local file access via attacker-controlled paths.

Disclosed 29 October 2026 · Record updated 29 October 2026

Impact

Attackers can use prompt injection to invoke arbitrary pandas functions, potentially executing code via pandas.read_pickle or reading local files through other readers.

Our coverage

A marionette puppet's strings tangle as invisible hands inject commands through its wooden mouth, forcing it to execute unpredictable jerking motions while reaching toward forbidden filing cabinets.

Vulnerabilities

Agno agent framework flaw lets prompts run arbitrary pandas calls

A vulnerability in Agno's PandasTools through version 3.1.11 lets injected prompts trigger code execution or local file reads, according to the entry published on 29 October 2026.

29 Oct 2026

Sources

nvd.nist.govhttps://nvd.nist.gov/vuln/detail/CVE-2026-119050