For teams building agent workflows, it helps to keep the risk review, input scanning, and run evidence as separate checks. That produces artefacts an operator can inspect instead of a single opaque pass/fail result.

Find them here: https://apify.com/zentrafoundry/agent-tool-risk-audit https://apify.com/zentrafoundry/prompt-injection-dataset-scanner-v2 https://apify.com/zentrafoundry/actor-flight-recorder-v2