linking-opportunities
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the
@openai/agent-browserpackage from NPM as a fallback browser tool. OpenAI is a well-known service and the package is a recognized utility for browser automation. - [COMMAND_EXECUTION]: The skill utilizes shell commands to install the browser tool (
npm i -g @openai/agent-browser) and subsequently executes CLI commands (agent-browser open,agent-browser find,agent-browser extract) to interact with website content. - [PROMPT_INJECTION]: The skill's core workflow involves ingesting and analyzing data from third-party websites, which creates a surface for indirect prompt injection attacks.
- Ingestion points: Data is retrieved from external prospect websites via the
agent-browsertool or Browser MCP during the validation and evidence capture phases. - Boundary markers: Absent. The instructions do not specify the use of delimiters or explicit 'ignore' prompts when handling external site content.
- Capability inventory: The agent possesses the capability to execute local shell commands and read project-specific files like
.seo-context.md. - Sanitization: There are no documented steps for sanitizing, escaping, or filtering the retrieved website content before the agent processes it for SEO analysis.
Audit Metadata