The Agent Skills Directory

[SAFE]: The skill implementation in akshare_tool.py and the instructions in SKILL.md were reviewed. No malicious code, obfuscation, or unauthorized data access patterns were found. The tool uses standard libraries for its stated purpose.- [INDIRECT_PROMPT_INJECTION]: While the skill ingests external data from financial APIs, it lacks exploitable capabilities, rendering the risk negligible.
Ingestion points: Financial data fetched via akshare API calls in akshare_tool.py.
Boundary markers: Data is structured into Markdown tables and headers for clear separation.
Capability inventory: The skill only performs read operations and returns text; it has no file system, network (beyond API calls), or execution privileges.
Sanitization: External data is processed by the pandas library and formatted as strings before being returned.

akshare