The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the user/agent to execute various shell commands including 'brew install', 'pip3 install', and 'sudo apt install'. These are standard for tool setup but involve executing commands with potentially elevated or system-level impact.
[PRIVILEGE_ESCALATION]: The installation instructions include 'sudo apt install python3-pip', which requires administrative privileges. While appropriate for the tool's installation, it is a high-privilege operation.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and manage task data ('todo.txt' format). There is a risk of indirect prompt injection if the agent reads tasks that contain malicious instructions (e.g., hidden in task descriptions, projects, or contexts) and subsequently executes them or allows them to influence its behavior.
Ingestion points: Reads from '~/todo.txt' and other configuration/data files specified in 'SKILL.md'.
Boundary markers: None identified in the instructions for separating task content from agent instructions.
Capability inventory: Uses 'topydo' CLI which can add, delete, and modify files. It also supports 'topydo edit' which may invoke external text editors.
Sanitization: No specific sanitization or escaping mechanisms for external task content are mentioned.
[COMMAND_EXECUTION]: The 'topydo edit' command allows opening tasks in a text editor, which could be exploited to execute arbitrary code depending on the environment's editor configuration.

topydo