team-performance-dashboard
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill's setup instructions suggest a manual process of cloning the vendor's repository and running a shell script (
sh install.sh). Although targeting the vendor's own repository, executing remote scripts is a significant capability. - [EXTERNAL_DOWNLOADS]: The skill recommends using the
skillsutility for installation and references companion tools from the vendor's domain and GitHub. These resources originate from the vendor and a trusted organization. - [PROMPT_INJECTION]: The skill processes untrusted lead replies and CRM data, presenting an indirect prompt injection surface.
- Ingestion points: Conversation messages from La Growth Machine, CRM records (HubSpot, Salesforce, Pipedrive), and local CSV/spreadsheet files.
- Boundary markers: The skill lacks explicit delimiters for data ingestion but instructs the agent to escape all user-provided text for artifact rendering.
- Capability inventory: Uses MCP tools to read campaign and lead data; writes snapshot files to the local
./.wtp/directory. - Sanitization: The instructions mandate HTML escaping of all user-supplied content to prevent injection in the rendered dashboard preview.
Audit Metadata