linkedin-personal-branding
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting untrusted data from external web pages. * Ingestion points: Profile data, activity, and posts are captured using the
mcp__chrome-devtools__take_snapshotandtake_screenshottools. * Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious instructions embedded within the LinkedIn profile content. * Capability inventory: The skill has the ability to click elements, navigate to new URLs, and hover over elements using MCP tools. * Sanitization: There is no evidence of sanitization or validation of the content extracted from the browser before it is processed by the agent. - [COMMAND_EXECUTION]: The skill utilizes several browser automation MCP tools including
mcp__chrome-devtools__list_pages,mcp__chrome-devtools__select_page,mcp__chrome-devtools__take_snapshot, andmcp__chrome-devtools__navigate_page. These are used appropriately to facilitate the profile analysis and metric calculations described in the core workflow. - [SAFE]: No evidence of hardcoded credentials, malicious obfuscation, persistence mechanisms, or unauthorized data exfiltration was found across the analyzed files.
Audit Metadata