The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements a 'Data-driven iteration' workflow in SKILL.md that ingests external data from CSV files, pasted text, and API exports. This creates a surface for indirect prompt injection as the agent is instructed to process this data without specific boundary markers or sanitization logic.
Ingestion points: SKILL.md (Mode two: Data-driven iteration).
Boundary markers: None identified in the instructions.
Capability inventory: Text generation, platform-specific schema application, and recommendation of external generation tools.
Sanitization: No escaping or validation of external content is specified.
[REMOTE_CODE_EXECUTION]: The reference guide references/generative-tools.md includes instructions for the user to execute npx create-video@latest, which facilitates the download and execution of remote packages. It also details the use of Remotion, a framework involving dynamic code execution of React and TypeScript to render video files.
[COMMAND_EXECUTION]: The documentation provides multiple shell commands for interacting with external APIs and setting up local software environments, including the use of curl, git clone, and make commands.
[EXTERNAL_DOWNLOADS]: The skill references numerous external AI platforms and tools, including ElevenLabs, Ideogram, Flux, and the Voicebox repository on GitHub. These represent dependencies on external infrastructure and codebases.

ad-copywriter