google-drive

The supplied code's sole behavior is adding two general AgentPMT skills. Although account setup could support a Drive workflow, this chunk contains no Drive-specific implementation or invocation connecting that setup to the declared purpose. The mismatch concerns the chunk's materially different primary behavior; the contents and behavior of the installed skills are not shown.

The recommended npx skills add command does not pin either the npm installer version or the source repository revision. Although this is a skill installer rather than an MCP server, following the recommendation could execute a compromised installer or import subsequently altered agent instructions.

This setup recommendation uses an unversioned npm executable and a mutable skill repository reference. Installation is conditional rather than automatic, which reduces immediate exposure, but following it introduces a supply-chain trust boundary.

The account-setup installation command leaves both the installer and downloaded instructions unpinned. Altered setup instructions could influence credential handling or connection configuration, making this dependency more sensitive than an informational skill.

The executable installation example runs npx skills without an exact version and retrieves skill content without an immutable revision. A compromised release or repository update could therefore change what executes locally or what instructions the agent later follows.

This command installs account/API setup instructions through an unpinned npm installer and mutable repository source. Compromise of either dependency could introduce local code execution or malicious credential-setup guidance, although no compromise is demonstrated here.

The reference repeats an installation command whose executable version and skill source revision are not fixed. Its placement in reference documentation lowers the likelihood of execution but does not remove the supply-chain exposure if followed.

This reference repeats the unpinned installation route for the credential-related setup skill. The risk is dependency substitution or compromise, not evidence of malicious behavior in the supplied document.

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

The explicit activation keywords include "copy file", "file id", and "new name", which commonly occur in local filesystem tasks and other services. Although the surrounding text describes Google Drive, these keywords are not explicitly conditioned on a Google Drive context and could cause unintended invocation.