todoist-automation
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches tool definitions and connects to the Rube MCP gateway at
rube.appto facilitate Todoist integration.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines workflows that ingest external data from Todoist tasks, creating a potential surface for indirect prompt injection through task content or descriptions.\n - Ingestion points:
TODOIST_GET_ALL_TASKS,TODOIST_GET_TASK, andTODOIST_GET_COMPLETED_TASKS_BY_COMPLETION_DATE(SKILL.md).\n - Boundary markers: None identified in the instructional workflows.\n
- Capability inventory: Includes task creation, modification, and deletion tools (
TODOIST_CREATE_TASK,TODOIST_UPDATE_TASK,TODOIST_DELETE_TASK).\n - Sanitization: No explicit data sanitization or escaping instructions are provided for processed task strings.
Audit Metadata