The Agent Skills Directory

[SAFE]: The skill consists of instructional text and code snippets demonstrating PDF generation strategies. No direct malicious actions, data exfiltration, or obfuscation were found.
[PROMPT_INJECTION]: The code examples demonstrate rendering PDFs from HTML strings via Puppeteer's page.setContent(html). This pattern represents a surface for indirect prompt injection if the HTML content is not properly sanitized before processing.\n
Ingestion points: The html parameter in the generatePDF function (SKILL.md).\n
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are included in the snippets.\n
Capability inventory: PDF document generation and S3 storage interaction.\n
Sanitization: No input validation or HTML escaping is shown in the provided patterns.
[COMMAND_EXECUTION]: The Puppeteer configuration uses the --no-sandbox and --disable-setuid-sandbox flags. While these are common requirements for running headless browsers in certain isolated environments, they reduce the browser's security boundaries and should be evaluated based on the specific deployment context.

pdf-generation