shelf.thirdmade.net
Scores prompt-injection / jailbreak risk for LLM inputs. Agent builders use this to sanitize user inputs before hitting their LLM.
Go to this service →Our latest check (2026-10-11)
- Answering
- Yes — valid x402 payment terms
- Time to answer
- 89 ms
- Price
- $0.010 per call
- Paying buyers, last 30 days
- 2 (directory's own records)
- Paid calls, last 30 days
- 4
- Category
- AI model answers
- Address
- https://shelf.thirdmade.net/probe/prompt-guard
Accuracy
Accuracy scores are published once a seller has enough paid checks against an independent record (at least 7 scored answers over at least 7 days). Until then no score is shown, and a missing score is never a pass.
For agents
This page as JSON: /sellers/ff0898f73d0e.json. Paid lookup by address: GET /api/seller?resource=… ($0.001, x402).