rivet-actors

Warn

Audited by Socket on Nov 8, 2026

5 alerts found:

Securityx3Anomalyx2
SecurityMEDIUM
examples/docs/actors-permissions/chat-room.ts

No clear malware behavior is present. The hardcoded, predictable authentication tokens are a significant security weakness if this actor is deployed with these values, especially because the admin token grants moderation permissions. Message rendering safety depends on downstream consumers.

Confidence: 99%Severity: 76%
SecurityMEDIUM
examples/docs/actors-permissions/caching-tokens.ts

The code has a serious authentication flaw: any non-empty token is accepted as the same hard-coded user. This can permit unauthorized connections and should not be used for real authentication without proper token verification. No evidence of malware or data exfiltration appears in this fragment.

Confidence: 99%Severity: 82%
SecurityMEDIUM
examples/docs/actors-permissions/create-conn-state.ts

No clear malware is present. The token validation is a significant authentication flaw: any nonempty token can connect with a fixed member identity and send messages. Message content is stored and broadcast without sanitization, so consumers should handle it safely.

Confidence: 99%Severity: 76%
AnomalyLOW
examples/docs/general-jwt/quickstart/server.ts

This code implements a token-issuing endpoint, not evident malware. Its main security concern is that it trusts the caller-controlled `x-demo-user` header as authentication, potentially allowing identity impersonation and unauthorized token issuance; the impact depends on deployment and the RivetKit token's permissions. The administrative token is read from `RIVET_ADMIN_TOKEN` and is not directly returned. Severity: medium security concern; no suspicious domains or hardcoded credentials are present.

Confidence: 98%Severity: 58%
AnomalyLOW
examples/docs/actors-request-handler/proxy.ts

This module is a generic request-forwarding gateway. The primary security concern is that user-controlled route parameters are used to (a) select an actor instance and (b) construct the destination URL via new URL(actorPath, "http://actor"), potentially allowing absolute URL overrides, and (c) proxy the entire incoming request (c.req.raw) into the forwarded Request. No validation/allowlisting is present in the shown code. While there is no clear evidence of intentional malware, the relay/target-injection pattern presents a credible SSRF/open-proxy and sensitive-header forwarding risk depending on deployment and rivetkit actor.fetch semantics.

Confidence: 62%Severity: 66%
Audit Metadata
Analyzed At
Nov 8, 2026, 12:21 AM
Package URL
pkg:socket/skills-sh/rivet-dev%2Fskills%2Frivet-actors%2F@9bca6b0fb3ca84312739d385fc66617e1e70f5d791c8d5b87b8ab10faaa83854