Original community demo · Video hosted on X
Workflow guardrails
A command-safety demo updates its risk warning as a shell command changes, distinguishing a build-folder cleanup from a much broader deletion.
Opens the author's original post in a new tab.
The implementation ideas below are our interpretation of the community demo. Check the original post for the author’s actual setup.
Provide an extra review signal before a human or an agent runs a command with potentially destructive effects.
Evaluate the proposed command with its working directory and stated intent. Display the risk result, then let a separate execution policy decide whether the action may proceed.
These examples explain a possible workflow; they are not recorded demo results or API request and response formats.
A probabilistic warning is not an execution sandbox. Shell expansion, indirect commands, and missing context can change the actual effect; enforce permissions outside the model.
Define the context and a small set of choices in the playground, then use the API guide to connect your application.
View demoA Rust desktop application classifies live network traffic to highlight potential anomalies as they arrive.
View demoA PostgreSQL experiment identifies personal-information spans and applies role-dependent masking while keeping the surrounding message readable.
View demoJev Ultrafast turns indexed page elements into candidate actions, demonstrated through a flight-search workflow.
