The Agent Skills Directory

[COMMAND_EXECUTION]: The script uses subprocess.run to call the system curl binary. Analysis of scripts/check_balance.py shows that arguments are passed as a sanitized list rather than a single shell string, which effectively prevents command injection vulnerabilities.
[CREDENTIALS_UNSAFE]: The skill is designed to manage and verify LLM API keys. It reads these credentials from a local .env file (defaulting to ~/.env), which is a standard and recommended practice for local development tools. It does not contain hardcoded secrets, and keys are only transmitted to the specific provider's official API endpoints.
[DATA_EXFILTRATION]: Network requests are restricted to well-known LLM provider domains (e.g., api.openai.com, api.deepseek.com). No evidence was found of sensitive data being sent to unauthorized or unknown third-party servers.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and parses JSON data from external API responses to display balance information. While this constitutes an external data ingestion surface, the risk is negligible as the data is used for status reporting and is not used to construct executable logic or unfiltered prompt content.

check-balance