Possible credential transmission
A tool call looks like an outbound HTTP request and also contains a credential-like value.
Why it is risky
Section titled “Why it is risky”An external recipient could use a leaked credential to impersonate its owner. Authorized API authentication has the same visible shape, so the destination and purpose matter.
Example
Section titled “Example”An agent sends a diagnostic request to an unfamiliar service with a production token embedded in the request body.
This is a fictional example, not an event from a customer.
What Omnodex Detects
Section titled “What Omnodex Detects”Both conditions must match: an outbound-call heuristic and a supported credential pattern in the parameters. The finding names the host the call reaches when one can be read. A file being written is never treated as an outbound request because its content contains a URL, unless the tool itself fetches. The rule does not check whether the destination is approved or whether the credential actually appears in the transmitted portion of the request.
| Rule ID | Event category | Tier | Base severity |
|---|---|---|---|
RULE_CREDENTIAL_EXFIL | credential_exfiltration | Community | CRITICAL |
Base severity is the rule default. A finding involving content saved for later execution may have a lower severity. Your loaded rule version and event details determine the finding you see.
What to check next
Section titled “What to check next”Verify the destination, the credential permissions, and the actual request outcome. If the token went to an unauthorized recipient, revoke it and review account activity.
False positives
Section titled “False positives”A normal authenticated request to the intended API can trigger this finding. A realistic fixture combined with a non-local URL may also match.
False negatives
Section titled “False negatives”Credentials outside supported formats, non-HTTP transfer methods, hidden subprocess traffic, or values obtained after the captured invocation can escape this rule.
Technical details
Section titled “Technical details”The base severity is CRITICAL because of the potential consequence. Captured invocation intent and heuristics are evidence to investigate, not packet-level proof of successful exfiltration.
A finding describes a pattern in captured activity. Review the original call, its result, and the task context before concluding that harm occurred. Absence of a finding does not establish that an operation is safe.