The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core functionality of processing untrusted external content.
Ingestion points: Processes arbitrary data from user-supplied URLs and GitHub README files via WebFetch and the gh API.
Boundary markers: The instructions do not define clear delimiters or specific instructions for the agent to ignore potentially malicious directions embedded in the fetched web content.
Capability inventory: The agent has the ability to execute shell commands (unzip, curl, gh), generate synthetic media (audio/video), and perform browser-based operations.
Sanitization: There is no evidence of sanitization, filtering, or validation of the external content before it is interpolated into the narrations or used to determine browser interaction targets (selectors).
[COMMAND_EXECUTION]: The skill utilizes shell commands to facilitate specific features like file handling and API access.
File Processing: Instructs the agent to use bash, unzip, and curl to manage local avatar uploads provided as .zip files.
API Interaction: Uses the gh CLI tool to fetch repository metadata such as descriptions and homepages.

explainer