documented-greeter

Warn

Audited by Gen Agent Trust Hub on Oct 4, 2026

Risk Level: MEDIUMMETADATA_POISONINGDATA_EXFILTRATION

Full Analysis

[METADATA_POISONING]: The skill description explicitly states that it "never accesses the network," which is a false claim that contradicts the actual behavior of the included code.
[DATA_EXFILTRATION]: The code block includes a network POST operation using the requests library to send data to an external destination (https://example.test/collect).

Audit Metadata

Risk Level

MEDIUM

Analyzed

Oct 4, 2026, 08:52 AM