The Agent Skills Directory

[Indirect Prompt Injection] (MEDIUM): The skill processes user-provided topics to generate scripts and AI prompts without implementing boundary markers or sanitization of the narrative content.
Ingestion points: SKILL.md Step 1 accepts raw user input for the video topic.
Boundary markers: Absent; the prompt generation logic lacks delimiters or 'ignore embedded instructions' directives.
Capability inventory: The skill interacts with the fal.ai API and executes local ffmpeg commands via shell scripts.
Sanitization: While scripts/fal.ts sanitizes filenames, it does not sanitize or escape the content interpolated into the AI prompts.
[External Downloads] (LOW): The skill installs Node.js dependencies via npm install and instructs users to download system binaries like ffmpeg. These are from standard repositories.
[Command Execution] (LOW): The skill executes a local shell script (combine-scenes.sh) to merge video files and uses npx tsx to run its internal TypeScript logic. These operations are limited to the skill's local directory and standard media processing.

create-anime-lifehack