registry / Credential exfiltration / @ricsam/r5d-browser / 0.0.190-prod.1790636572228.gbad0936ff3327d40d4da94c6be11353f3bc8aad0
@ricsam/r5d-browser@0.0.190-prod.1790636572228.gbad0936ff3327d40d4da94c6be11353f3bc8aad0
`r5d-browser start --auto-update` enables automatic client updates for this environment and state directory. The setting persists; `--no-auto-update` disables it. The client checks the authenticated environment's selected release once a minute and stages
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThe configurable destination is not host-bound before credential forwarding, creating a concrete conditional exposure risk. The inspected code otherwise shows user-invoked browser automation and no lifecycle-based or hidden exfiltration behavior.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
Running the CLI start command with a configured origin and token.
Impact
A token intended for one service could be exposed if an operator configures an untrusted HTTPS origin.
Mechanism
The relay derives an origin from options, environment, or config and sends the token in an Authorization header to that origin.
AI rationale
The configurable destination is not host-bound before credential forwarding, creating a concrete conditional exposure risk. The inspected code otherwise shows user-invoked browser automation and no lifecycle-based or hidden exfiltration behavior.
Evidence
dist/mjs/personal-main.mjs
Network endpoints1
r5d.dev
Decision evidence
public snapshotAI called this Suspicious at 86.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- Package metadata selects a JavaScript module entrypoint.