The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing local scripts and system commands to perform its core functions:
Executes python scripts/extract-pptx.py to extract content from user-provided PowerPoint files.
Runs bash scripts/deploy.sh to bundle and upload presentation files to the Vercel hosting platform.
Runs bash scripts/export-pdf.sh to capture slide screenshots and generate a PDF document.
Uses the open command to launch generated HTML files in the default system browser.
[EXTERNAL_DOWNLOADS]: The skill manages dependencies by downloading and installing external packages and binaries at runtime:
Installs the python-pptx library using the Python package manager (pip).
Uses npx to download and execute the Vercel CLI for deployment operations.
Uses npx to install the Playwright browser automation tool and its associated Chromium binaries (~150MB).
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, creating a potential surface for indirect prompt injection:
Ingestion points: Processes user-provided .pptx files during conversion (Phase 4) and reads existing HTML presentations for enhancement (Phase 0).
Boundary markers: There are no explicit instructions to use delimiters or ignore instructions embedded within the processed presentation content.
Capability inventory: The skill has access to the filesystem, can execute shell commands, and performs network operations during deployment.
Sanitization: The instructions do not define any sanitization or validation routines for content extracted from external files before it is interpolated into new presentation code.
[DYNAMIC_EXECUTION]: The skill dynamically generates functional HTML files containing inline CSS and JavaScript based on user input and pre-defined style templates. This generated code is subsequently executed when opened in a browser or processed by the Playwright renderer.

frontend-slides-frontend-slides