The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests external financial data from a CLI tool, which is a potential surface for indirect prompt injection from remote API responses. The instructions mitigate this by directing the agent to use structured JSON schemas and to save large datasets to local files rather than including them directly in the conversation context.- [DATA_EXFILTRATION]: The skill documentation includes the --output parameter for the CLI tool, which allows the agent to write query results to local files. This is a standard and expected feature for managing financial data exports.- [CREDENTIALS_UNSAFE]: The skill mentions the use of authentication via the HITHINK_FINANCE_API_KEY environment variable or the --api-key-stdin flag. The instructions correctly point to using system credentials or environment variables rather than hardcoding secrets.

hithink-finance-futures