sevdesk-automation
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
- Indirect Prompt Injection (LOW): The skill is designed to ingest and execute tool schemas fetched dynamically from an external search tool.
- Ingestion points: Tool definitions and schemas returned by
RUBE_SEARCH_TOOLSas described in Step 1 of the Core Workflow. - Boundary markers: None. The instructions do not provide delimiters or warnings to ignore malicious instructions embedded in the returned schemas.
- Capability inventory: The agent is granted capabilities to execute business operations via
RUBE_MULTI_EXECUTE_TOOLandRUBE_REMOTE_WORKBENCH. - Sanitization: Absent. The skill assumes the integrity of the tool slugs and arguments provided by the external search service.
- External Downloads (LOW): The skill references an external MCP server endpoint (
https://rube.app/mcp) that is not part of the trusted repository or organization list. While standard for Rube/Composio integrations, it constitutes an external dependency. - No Local Code (SAFE): The skill consists entirely of markdown instructions and does not bundle any Python, JavaScript, or shell scripts, reducing the risk of direct local exploit execution.
Audit Metadata