The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected. The skill's code and instructions are consistent with its stated purpose as a service integration.
[COMMAND_EXECUTION]: The skill utilizes several Bash scripts to interface with the Pexo API. These scripts are transparent, well-documented, and use standard utilities to perform their tasks.
[DATA_EXFILTRATION]: Communication is directed to the vendor's API at https://pexo.ai. The skill handles user-provided API keys via a local configuration file (~/.pexo/config) and includes mechanisms to verify connectivity and authentication safely.
[PROMPT_INJECTION]: The skill instructs the agent to relay user messages verbatim to the backend service, which creates a surface for indirect prompt injection.
Ingestion points: User-provided product descriptions and video instructions are passed directly to the pexo-chat.sh script.
Boundary markers: The instructions do not define specific delimiters or "ignore previous instructions" guards for the relayed content.
Capability inventory: The skill can execute local scripts that perform network operations and read/write to the skill's local state directory.
Sanitization: The scripts perform local syntax validation for asset tags but do not filter or sanitize the natural language content of user messages.

product-video