The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The documentation file references/ifind-token-management.md contains hardcoded authentication credentials for the iFinD service, including a specific username (htzqywb001) and password (htzqywb888).\n- [PERSISTENCE]: The file references/ifind-token-management.md provides instructions for establishing persistence by configuring a cron job (0 10 * * 1,4 ...) to execute the skill's Python script periodically for automated token management.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes data from various untrusted external sources, including financial APIs and PDF reports, creating an attack surface for indirect prompt injection.\n
Ingestion points: Data retrieved from external APIs (iFinD, AKShare, EastMoney) and content extracted from downloaded corporate announcement PDFs as described in references/announcement-deep-reading.md.\n
Boundary markers: None identified in the provided documentation for separating external data from agent instructions.\n
Capability inventory: The skill possesses capabilities for network access (requests) and file processing (fitz/PyMuPDF) through the stock_api.py and cninfo_download.py scripts.\n
Sanitization: No specific sanitization or filtering mechanisms are described for the external content before it is processed by the agent.

tonghuashun