The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates by executing shell commands using the kiwoomcli utility. It includes pre-flight environment checks to verify the tool's path and execution status within the current session.
[EXTERNAL_DOWNLOADS]: The documentation directs the user to install the kwcli package using uv. This is a standard dependency installation process for the required tooling.
[DATA_EXFILTRATION]: While the skill retrieves sensitive financial information such as holdings and account status, it mitigates risk by explicitly forbidding the logging or display of raw authentication tokens and account numbers.
[CREDENTIALS_UNSAFE]: The skill follows security best practices by requiring secrets to be stored in the OS-level keyring via the CLI's setup command and specifically prohibits the use of plaintext .env files.
[PROMPT_INJECTION]: The skill establishes operational guardrails, including a mandatory two-stage order process (preview then confirm) and default restrictions to mock investment profiles.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface. Ingestion points: Financial and market data is ingested from kiwoomcli JSON output as defined in SKILL.md. Boundary markers: Absent; there are no explicit delimiters or warnings provided to the agent to ignore instructions embedded in the ingested data. Capability inventory: The skill maintains the ability to execute kiwoomcli shell commands. Sanitization: No validation or escaping of the external data is performed before it is processed by the agent.

kiwoom-broker