The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to clone a repository from an untrusted GitHub account (https://github.com/plushyta/Awesome-Marketing-Skills.git).
[PRIVILEGE_ESCALATION]: The troubleshooting section suggests executing sudo apt-get install, which involves acquiring root privileges on the host system to install software.
[DYNAMIC_EXECUTION]: The Puppeteer implementation uses the --no-sandbox and --disable-setuid-sandbox flags. Disabling the browser sandbox is a dangerous practice that allows a compromised browser process to access the host system with the same privileges as the parent process.
[INDIRECT_PROMPT_INJECTION]: The skill generates brochures by ingesting content from external research tools (Tavily and Firecrawl) without evident sanitization, creating a surface for indirect prompt injection.
Ingestion points: Research data fetched via Tavily and Firecrawl MCP servers in the researchLocation function.
Boundary markers: The skill lacks explicit delimiters or instructions for the agent to ignore potentially malicious content within the research data.
Capability inventory: The skill has capabilities for network operations, file writing (PDF generation), and shell execution (via git and uvx).
Sanitization: No sanitization or escaping is performed on the external research results before they are rendered into HTML for PDF export.
[COMMAND_EXECUTION]: The skill uses uvx to execute various MCP servers and git clone to retrieve the library, both of which involve running shell commands based on external references.

awesome-marketing-skills-library