The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions for the agent to install the finance-cli tool by executing a script directly from a remote URL via curl -fsSL https://raw.githubusercontent.com/kelvin6365/finance-cli/main/install.sh | bash. This represents unverified remote code execution from a third-party source.
[COMMAND_EXECUTION]: The installation workflow in references/install.md guides the agent to perform environment persistence by modifying shell configuration files (~/.zshrc or ~/.bashrc). It also instructs the agent to use sudo for installation to system-protected directories like /usr/local/bin if permissions are insufficient.
[EXTERNAL_DOWNLOADS]: The skill recommends installing software from unverified third-party sources using commands like bun add -g github:kelvin6365/finance-cli and npx skills add kelvin6365/finance-cli-skill.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) due to its dependency on external CLI data.
Ingestion points: Data returned from finance schema, finance status, and transaction listings.
Boundary markers: The instructions lack delimiters or specific warnings to ignore instructions embedded in the processed financial data.
Capability inventory: The agent has extensive capabilities, including executing shell commands, installing scripts, and writing to files.
Sanitization: No input validation or output sanitization of the CLI results is specified in the skill body.

finance-cli