The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes dynamic output from external command executions to drive its internal logic, which represents an injection surface.
Ingestion points: Output from circle wallet login --init (specifically the request ID) and circle terms show --init --output json (specifically the terms of use URL, privacy policy URL, and terms notice text).
Boundary markers: The skill does not implement explicit boundary delimiters or special instructions to isolate external tool output from the agent's control context.
Capability inventory: The skill has the capability to perform network-based authentication, execute wallet creation commands, and record legal agreement acceptance through the circle CLI.
Sanitization: No explicit sanitization, filtering, or validation is performed on the live values returned by the CLI before they are presented to the user or used in subsequent steps.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to install the @circle-fin/cli package from the npm registry. This is a standard installation procedure for a tool provided by a well-known fintech service.
[COMMAND_EXECUTION]: The skill uses various local shell commands (e.g., circle wallet, circle terms) to interact with the wallet infrastructure. These operations are essential to the primary function of the skill.

use-agent-wallet