The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions and code examples that involve executing shell commands. Specifically, in rules/avatars.md, it demonstrates the use of child_process.execSync to run the open command on macOS, and provides terminal examples using xdg-open (Linux) and start (Windows) to open URLs in a web browser.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and process data from the HeyGen API (api.heygen.com), creating a potential surface for indirect prompt injection where malicious content from the API response could influence agent behavior.
Ingestion points: Data is ingested from HeyGen API endpoints such as /v2/avatars, /v2/voices, and /v2/templates as described in multiple files including rules/avatars.md and rules/voices.md.
Boundary markers: The provided code snippets do not include explicit boundary markers or warnings to the agent to ignore instructions embedded in the API data.
Capability inventory: The skill utilizes network operations (fetch, requests, httpx), file system writes (fs.writeFileSync in rules/video-status.md), and shell command execution (execSync in rules/avatars.md).
Sanitization: There is no evidence of sanitization or strict validation of string data returned from the API before it is used in subsequent operations or prompt construction.

heygen-best-practices