The Agent Skills Directory

[COMMAND_EXECUTION]: The 'Data Attributes Reference' table contains an unexpected string that appears to be the output of a shell id command (uid=501(salmanqureshi) gid=20(staff)...) in place of a valid technical attribute. This exposure of local system details indicates a lack of verification in the documentation process.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install external code via npx skills add heygen-com/hyperframes. The package source is not identified as a trusted vendor, and the naming convention differs from official branding, which may lead to the installation of unverified software.
[REMOTE_CODE_EXECUTION]: The production workflow encourages dynamic code execution through npx and npm commands, such as adding transitions (npx hyperframes add flash-through-white) or initializing projects. This creates a surface for remote code execution from unverified package registries.
[COMMAND_EXECUTION]: Multiple instructions guide the agent to execute shell commands with broad environment access for rendering (npx hyperframes render), linting, and system diagnostics (npx hyperframes doctor), which could be exploited if the underlying scripts are compromised.

hyperframes