The Agent Skills Directory

[DATA_EXFILTRATION]: Accesses sensitive financial data including account balances, stock holdings, and profit/loss statements. It reads user credentials from ~/.finhay/credentials/.env to authenticate requests to the vendor's API.
[COMMAND_EXECUTION]: Executes shell scripts ./_shared/scripts/infer-sub-account.sh and ./_shared/scripts/request.sh for configuration management and API request signing. These scripts are referenced in the instructions but the source code was not included in the skill package, preventing verification of their security.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from the Finhay API (such as stock symbols, order history, and market news) which is then rendered in the agent's context.
Ingestion points: API endpoints defined in references/endpoints/portfolio.md and references/endpoints/orders.md.
Boundary markers: Absent; no delimiters or instructions are used to distinguish API data from agent instructions.
Capability inventory: Capability to execute shell scripts and read local files.
Sanitization: No input validation or sanitization of data retrieved from the trading endpoints is specified in the documentation.

finhay-portfolio