The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from various external sources which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill ingests data from external web pages via web_fetch and web_search, and from user-controlled databases via query_notion_database and search_notion. It also reads local event logs and memory files.
Boundary markers: There are no explicit boundary markers or instructions provided to the agent to treat content from these external sources as data only and to ignore any embedded commands.
Capability inventory: The skill has significant capabilities, including the ability to modify core planning files (stream_edit), create or modify automated recurring tasks (recurring_add, recurring_update_task), and manage scheduled tasks.
Sanitization: The skill lacks explicit sanitization of external inputs, although it does implement a '3+ Rule' and 'Conservatism Principle' that require multiple pieces of evidence or explicit user requests before taking action, which serves as a logical mitigation.

month-planner