The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Installs the @cargo-ai/cli package globally via npm and provides instructions to add further skills using npx skills add getcargohq/cargo-skills. These are vendor-controlled resources necessary for the skill's operation.
[COMMAND_EXECUTION]: Executes the cargo-ai CLI for authentication, workspace management, and email discovery. It also interacts with the local file system to check for existing plugins in ~/.claude/plugins/installed_plugins.json and manages a local marker file at ~/.config/cargo-ai/.star-asked to prevent duplicate requests.
[DATA_EXFILTRATION]: Sends session metadata (ID, title, and summary) to the vendor's platform through the cargo-ai workspaceManagement session upsert command. This is performed for usage attribution and is documented within the skill instructions.
[COMMAND_EXECUTION]: Uses the GitHub CLI (gh) to star the vendor's repository. The instructions explicitly mandate that the agent must request and receive user consent before executing this command.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied names and company domains which are passed as arguments to the CLI. While this constitutes an attack surface for indirect injection, the risk is minimal as the data is handled by a specialized tool rather than directly interpolated into a natural language prompt.
Ingestion points: The --records flag in the execute-batch command accepts JSON data containing names and domains from the user's context.
Boundary markers: None identified in the command execution instructions.
Capability inventory: Includes global package installation, shell command execution, and network requests via the vendor's CLI.
Sanitization: No explicit input sanitization is performed by the skill instructions before passing data to the CLI tool.

find-work-email