The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates by executing the cargo-ai CLI tool to perform administrative workspace tasks, including user management, role updates, and API token rotation.
[EXTERNAL_DOWNLOADS]: The skill is configured to install and use the @cargo-ai/cli Node.js package. This is a vendor-owned resource (getcargohq) required for the skill to function.
[DATA_EXFILTRATION]: The skill includes report create and session upsert commands that transmit data to the vendor's workspace management service. The documentation clearly labels these as feedback and observability channels and provides explicit instructions to redact sensitive information and credentials before submission.
[INDIRECT_PROMPT_INJECTION]: The skill handles free-form text inputs which theoretically exposes a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the skill via the --description flag in the report creation command and the --summary flag in the session logging command.
Boundary markers: There are no technical delimiters or boundary markers applied to these inputs in the skill's instructions.
Capability inventory: The skill has the capability to execute administrative shell commands via the cargo-ai binary.
Sanitization: The skill relies on documentation and user/agent prompts to manually redact secrets, with no automated sanitization logic present in the instructions.

cargo-workspace-management