experience-cms-brand-create
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes OS-specific commands (
open,xdg-open, andStart-Process) to trigger thevscode://salesforce.salesforcedx-vscode-ui-preview/open-brand-managerprotocol handler. This action is used to launch the Brand Toolkit UI panel in VS Code. The commands use hardcoded URI strings and do not interpolate user input, mitigating command injection risks. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input, such as brand names and design descriptions, to populate a JSON template.
- Ingestion points: User-provided brand intent, name, and color descriptions in Step 1.
- Boundary markers: The instructions do not define explicit delimiters for user input, but they restrict the agent to modifying only specific "leaf values" within a predefined JSON structure.
- Capability inventory: The skill has the capability to write files to the
.digitalExperience/brands/directory and execute shell commands to open the VS Code protocol handler. - Sanitization: The skill implements a specific regex-based slugification rule (
title.toLowerCase().replace(/[^a-z0-9]+/g, '_').replace(/^_|_$/g, '')) for generating filenames, which prevents directory traversal or unauthorized file naming via the brand title.
Audit Metadata