qa
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references and executes
@team-attention/chromuxfrom the npm registry using thenpxcommand to facilitate browser-mode testing. This is a vendor-owned package associated with the skill's author ('team-attention') and follows standard development practices for remote dependency usage. - [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection due to its core functionality of processing untrusted data from applications-under-test.
- Ingestion points: The skill retrieves data from external environments via accessibility trees (
chromux snapshot), visual screenshots (mcp__computer-use__screenshot), and terminal screen buffers (tmux capture-pane). - Boundary markers: The instructions lack explicit delimiters or specific warnings to ignore instructions embedded within the data retrieved from target applications.
- Capability inventory: The skill has access to sensitive tools including
Bash(shell execution),Write/Edit(file system modification), and the ability to performgit commitoperations on the local repository. - Sanitization: There is no evidence of sanitization or validation of the ingested application data before it is processed by the agent.
Audit Metadata