monte-carlo-storage-cost-analysis

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses specific MCP tools (analyze_storage_costs, get_asset_lineage) to perform its primary function. These tools are designed to interact with Monte Carlo's data observability platform for warehouse analysis.
  • [SAFE]: Explicit instructions are included to prevent the agent from executing destructive actions, such as DROP TABLE, limiting the skill's scope to analysis and recommendations only.
  • [SAFE]: The 'Monte Carlo tool routing' section ensures that the agent uses the correct environment and credentials by specifying a bundled server path, which is a best practice for scoping tool access and avoiding credential confusion.
  • [SAFE]: The skill incorporates clear boundaries for presenting tool output, instructing the agent to present specific pre-formatted blocks (<!-- PRESENT_AS_IS -->) while excluding detailed categories until requested. This reduces the risk of processing unintended instructions from external data sources.
  • [SAFE]: The skill manages indirect prompt injection risks by using explicit output markers and instructions for verbatim presentation. Ingestion points include the analyze_storage_costs tool output, which is scoped by <!-- PRESENT_AS_IS --> and <!-- CATEGORY_DETAILS --> markers. The capability inventory is limited to analysis tools and the Read tool, with instructions to preserve formatting and linked MCONs without further transformation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 12:25 AM
Security Audit — agent-trust-hub — monte-carlo-storage-cost-analysis