The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The file scripts/get_data.py contains a hardcoded API credential (em_fjFqd4YB6Cqs52LF48XWbMDdLNq6MyNg) assigned as the default value for the EM_API_KEY environment variable.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it fetches and displays content from an external API without performing sanitization or using boundary markers.
Ingestion points: Data is ingested from the Eastmoney stock analysis API in scripts/get_data.py.
Boundary markers: The skill does not use delimiters or provide instructions to the agent to treat the API output as untrusted data.
Capability inventory: The script scripts/get_data.py possesses the capability to write files to the local filesystem using Path.write_text.
Sanitization: There is no evidence of filtering or validation of the displayData or other fields returned by the API before they are presented to the agent.
Remediation: Use clear boundary markers (e.g., XML tags or unique delimiters) when interpolating API data into prompts and explicitly instruct the agent to ignore any commands contained within that content.
[SAFE]: The skill transmits user queries to the Eastmoney API (https://ai-saas.eastmoney.com/proxy/app-robo-advisor-api/assistant/stock-analysis) to generate financial reports, which is the primary intended function of the skill.

stock-diagnosis