The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill connects to a third-party service at api.labnana.com to generate slide images using the gpt-image-2 model. This requires the user to provide their own API key via a .env file.
[COMMAND_EXECUTION]: The skill executes local TypeScript scripts (merge-to-pptx.ts, merge-to-pdf.ts) and a Python script (generate_slide.py) to manage file processing and API communication. There is a potential risk of command injection in the shell template used to call generate_slide.py if LLM-generated content is not properly escaped.
[PROMPT_INJECTION]: The skill processes untrusted user-supplied markdown files (path/to/content.md) to generate slide outlines and image prompts, making it susceptible to indirect prompt injection.
Ingestion points: The source markdown file analyzed in Step 1.2.
Boundary markers: No explicit delimiters or boundary instructions are mentioned to protect the agent from instructions embedded within the source content.
Capability inventory: File system write access, network communication (API), and shell command execution.
Sanitization: No sanitization or filtering logic is described for the input content prior to outline generation.
[DATA_EXPOSURE]: The skill reads user configuration from the user's home directory ($HOME/.baoyu-skills/baoyu-slide-deck/EXTEND.md) to load style and audience preferences.

baoyu-slide-deck