todoist
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from Todoist, including tasks, projects, comments, and reminders, which may contain malicious instructions designed to manipulate the agent.
- Ingestion points: The agent retrieves data through the
todoist list-toolsandtodoist call-toolcommands, which can return user-defined content from the Todoist API. - Boundary markers: The instructions do not specify any delimiters or warnings for the agent to distinguish between platform instructions and data retrieved from Todoist.
- Capability inventory: The skill possesses significant capabilities, including the ability to create, modify, and delete tasks and projects via
todoist call-tool. - Sanitization: There is no evidence of sanitization or filtering of the content retrieved from Todoist before it is processed by the agent.
Audit Metadata