The Agent Skills Directory

[COMMAND_EXECUTION]: Executes local scripts included with the skill repository to perform core functions.
Evidence includes the use of scripts/extract-pptx.py for PowerPoint conversion and scripts/deploy.sh for Vercel deployment.
The skill uses scripts/export-pdf.sh to generate static PDF versions of presentations.
[EXTERNAL_DOWNLOADS]: Installs necessary third-party dependencies from established package registries.
Detects the installation of the python-pptx library via pip.
Downloads the Chromium browser binary using Playwright (npx playwright install chromium) to facilitate PDF exports.
Utilizes the Vercel CLI (npx vercel) for web hosting tasks.
[EXTERNAL_DOWNLOADS]: Fetches visual templates, CSS configurations, and design guidelines from the project's GitHub repository.
References STYLE_PRESETS.md, viewport-base.css, and various template files hosted at github.com/zarazhangrui/frontend-slides.
[DATA_EXFILTRATION]: Provides a functional path for users to upload and host their generated presentations on Vercel.
This behavior is transparently described as a sharing feature and targets a well-known cloud hosting service.
[PROMPT_INJECTION]: The skill processes untrusted external data (PowerPoint files and text notes) to generate executable HTML/JS presentations, which constitutes a potential surface for indirect injection.
Ingestion points: External .pptx files and user-provided notes processed in Phase 1 and Phase 4.
Boundary markers: The skill instructions do not specify delimiters or warnings to ignore instructions embedded within the user data.
Capability inventory: The skill has the capability to execute shell/Python scripts and perform network operations via the Vercel CLI.
Sanitization: There is no explicit documentation of sanitization or filtering logic applied to the input content before it is interpolated into the HTML presentation code.

frontend-slides-frontend-slides