The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes and stores user-supplied text (todo titles) and presents them back to the agent, creating a surface where malicious instructions embedded in a task could attempt to influence the agent's future behavior.
Ingestion points: The addTodo API in apis/addTodo.js accepts a user-provided title parameter which is saved to the todo_items database collection.
Boundary markers: The getTodoList API and the todo-list-card component do not use explicit delimiters or specific instructions to the agent to treat retrieved task titles as data only and ignore any embedded commands.
Capability inventory: The skill provides write and delete capabilities via toggleTodo and deleteTodo APIs, which could be targeted by instructions hidden in data.
Sanitization: The skill does not perform sanitization, filtering, or validation on the title field beyond checking that it is not empty.

todolist-skill