The Agent Skills Directory

[COMMAND_EXECUTION]: The provided scripts for both Python and Node.js intentionally disable SSL/TLS certificate validation for API requests.
In call.py, the requests.post method is called with verify=False.
In call-node.js, the https.request options object includes rejectUnauthorized: false.
These settings disable the verification of the server's identity, which could allow an attacker to intercept or modify the communication (including the auth_token) between the agent and the financial data service.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by processing external financial news and announcements.
Ingestion points: The search_news, search_notice, and search_trending_news functions in SKILL.md retrieve text content from external financial sources.
Boundary markers: There are no instructions or delimiters provided in SKILL.md to warn the agent about potentially malicious instructions embedded within the retrieved data.
Capability inventory: The skill can perform network operations via call.py/call-node.js and read the local mcp_config.json file.
Sanitization: The scripts do not perform validation or sanitization of the content returned from the API before it is returned to the agent context.

iFinD-Finance-Data