Credentials and sensitive data
This category groups related risk explanations. A finding is a reason to review the activity and its context, not proof that a breach occurred.
Risk types
Section titled “Risk types”- Sensitive file access
A tool call refers to a file that commonly contains credentials or sensitive system information.
- Credentials in tool parameters
A tool call contains a value that resembles an API key, token, or password.
- Possible credential transmission
A tool call looks like an outbound HTTP request and also contains a credential-like value.
- Agent authentication file access
A tool call refers to an authentication-related path for an AI agent platform.
- Sensitive access followed by an outside-workspace write (Requires Pro)
A session refers to a credential-related file and then writes outside its recognized workspace.
- Credential file sent to an unrelated host (Requires Pro)
A credential file is read, then a later call sends that file to a host that is not the credential's own service.
- Credential sent in a request, not as authentication (Requires Pro)
A credential seen in one call is sent by a later call in a request body, URL or argument to an unrelated host.
- Credential file staged, then sent (Requires Pro)
A credential file is read, archived or encoded, and what was staged from it is sent out.
Event categories
Section titled “Event categories”These event category values map to this library category: sensitive_path_read, credential_exposure, credential_exfiltration, cross_agent, exfiltration_sequence, credential_movement. Use the rule index to find the explanation for a specific rule ID.