The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to clone the explainer-studio repository from GitHub (https://github.com/yahyabedirhan/explainer-studio) if it is not already present on the local system.
[COMMAND_EXECUTION]: The skill executes various shell commands including git clone, npm install (via setup instructions), and specific project scripts such as npm run --silent root and npm run dev to manage video rendering and previews.
[REMOTE_CODE_EXECUTION]: The skill executes code and setup procedures downloaded from an external repository. It also adopts instructions from an external file (AGENTS.md) within that repository, which are permitted to override the skill's own rules.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for instruction injection by design, as it is instructed to read and prioritize instructions from an external file (AGENTS.md) located in the cloned repository.
Ingestion points: The AGENTS.md file located within the cloned explainer-studio directory (Flow Step 1).
Boundary markers: None specified; the skill explicitly states that the external file's rules 'win where it differs from this skill'.
Capability inventory: The agent has the capability to execute shell commands (npm, git), read project files, and write rendered video files to the local filesystem.
Sanitization: No explicit sanitization or validation of the instructions within AGENTS.md is performed before execution.

make-explainer