The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill contains a workflow for converting PowerPoint (.pptx) files into HTML presentations, which involves reading external, potentially untrusted data that is then processed by the agent.\n
Ingestion points: The scripts/extract-pptx.py script extracts text and images from user-supplied .pptx files into a JSON format.\n
Boundary markers: The instructions do not specify the use of clear delimiters or guardrail prompts when the extracted content is provided back to the agent for styling.\n
Capability inventory: The skill possesses the ability to execute shell commands (bash, python, open), write to the local filesystem, and perform network requests/package installations via Playwright.\n
Sanitization: There is no evidence of sanitization or filtering of the extracted text content to prevent embedded instructions from affecting the agent's behavior.\n- [DYNAMIC_EXECUTION]: The scripts/export-pdf.sh utility script generates a temporary Node.js script (export-slides.mjs) from an internal template and executes it to automate the PDF generation process.\n- [EXTERNAL_DOWNLOADS]: The PDF export functionality fetches and installs the playwright package and the Chromium browser binary from the official NPM registry and Microsoft's distribution servers.\n- [COMMAND_EXECUTION]: The skill uses various system commands to perform its tasks, including python3 for content extraction, bash for PDF exporting, and OS-specific commands (open, xdg-open, start) to launch the generated presentations in a web browser.

frontend-slides