verify-ui-change
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches documentation and technical specifications from the vendor's official domain (
docs.reticle.sh) usingcurlcommands. These resources provide context for the agent to use specific tools and predicates. - [REMOTE_CODE_EXECUTION]: Instructions include running
npx @reticlehq/server@latest initto install and configure the necessary SDK and server. This execution of a remote package from the npm registry is a standard part of the setup for the vendor's tooling. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process real-time data from a running web application, including console logs, network request metadata, and DOM element properties via tools like
reticle_act_and_waitandreticle_look. This untrusted data from the application under test creates a surface for indirect prompt injection, as the agent may act upon instructions embedded within the app's output. The skill does not explicitly define sanitization routines or strict boundary markers for this external data.
Audit Metadata