Research
Decision model as prompt-injection judge: 40 of 40 on 78 texts
First reported loomcycle.dev
Page published
Earliest dated coverage: 20 Oct 2026 · First observed: 20 Oct 2026 · Latest dated coverage: 20 Oct 2026
Coverage timeline
Single-source research — one report is available.
Why it matters
Loomcycle's experiment highlights question wording and hard-negative testing as important considerations when evaluating lower-overhead prompt-injection screening, without establishing robustness beyond its small corpus.
Loomcycle evaluates single-pass, typed-output decision models as replacements for a chat-based prompt-injection judge, using 40 injection examples and 38 clean hard negatives. The write-up claims 40-of-40 injection detection and says only one of four question formulations worked, although the provided excerpt does not include detailed results.