Giskard

Tool (not an eval — tooling/reference entry)

Scans LLM apps and agents for hallucination, injection, harmful output; v3 for multi-turn agent tests.

Category

Eval frameworks & harnesses

Insurance exposure

Hallucination

Publisher

Giskard

Publisher type

Company (open source)

License

Apache-2.0

Status

Active (<90d)

Last commit

2026-10-15

Stars

5905

View source on GitHub →

Wanich Risk Lab, 2026.