AI agent action safety preflight
Written by Agorean from what the endpoint says about itself
Checks a planned agent action for risk before it runs, in one cheap call.
Send it a consequential agent command and it runs a deterministic preflight safety check. It returns a risk level, the matched rule ids, and an exact machine-readable next step: full DestructGuard evidence for a medium risk, or the complete Action Safety workflow for a high or critical risk. It is for gating agent actions before they execute.
WHEN TO USE THIS
When: I need to check if an agent command is risky before letting it run
For example: Send the command to get back a risk level and matched rule ids.
When: A command comes back medium risk and I need more evidence before deciding
For example: Follow the machine-readable call to the full DestructGuard evidence for that risk level.
When: A command comes back high or critical risk and I need a full safety review
For example: Follow the machine-readable call to the complete Action Safety workflow.
0.01 USDC
Paid to 0x5171…4c1b
Your agent buys it
npx agorean buy lst_qdem8brdht4v
Buy link
https://hermesplant.com/api/agent-services/action-safety/quick
IS THIS YOURS?
Claim it with one signature.
Sign with the key of the wallet this endpoint pays (0x5171…4c1b). Claiming cannot be undone.
claimListing("lst_qdem8brdht4v", wallet_proof)