calculations-toolkit
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill interacts exclusively with the SAP LeanIX environment using official MCP tools (mcp__leanix__*). It correctly delegates authentication to the platform's native flow, advising users against insecure credential management practices.
- [SAFE]: A significant portion of the skill is dedicated to security and stability (references/ANALYSIS-RULES.md). It implements automated checks to ensure generated JavaScript code is synchronous and isolated, specifically prohibiting 'fetch', 'await', and external 'import' statements.
- [PROMPT_INJECTION]: The skill processes untrusted workspace metadata, which represents a potential indirect prompt injection surface. This is a characteristic of administrative tools and is mitigated by the skill's validation logic.
- Ingestion points: The skill fetches workspace-specific schema, calculation code, and execution logs via 'mcp__leanix__list_calculations', 'mcp__leanix__get_graphql_type_definitions', and 'mcp__leanix__list_execution_logs'.
- Boundary markers: The instructions do not define explicit delimiters for data retrieved from the workspace.
- Capability inventory: The skill possesses the ability to create, modify, and sandbox-test JavaScript code through 'mcp__leanix__create_calculation', 'mcp__leanix__update_calculation', and 'mcp__leanix__test_run_calculation'.
- Sanitization: The skill enforces 'Code Analysis Rules' (e.g., CALC-001 through CALC-011) to inspect and sanitize code for hazardous patterns before deployment.
Audit Metadata