boards-sprint-metrics
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
curlcommands to interact with Azure DevOps services. These commands are constructed using environment variables for organization names and personal access tokens, and include specific handling for PowerShell environments. - [EXTERNAL_DOWNLOADS]: The skill fetches configuration and metric data from
dev.azure.comandanalytics.dev.azure.com. These are well-known Microsoft services for development and analytics. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection because it ingests and processes data from external work items.
- Ingestion points: The skill retrieves work item attributes, iteration names, and titles from the Azure DevOps Analytics OData endpoint (SKILL.md).
- Boundary markers: Results are presented in structured markdown tables, though no specific delimiters or warnings are used to prevent the agent from interpreting work item content as instructions.
- Capability inventory: The skill uses
curlto perform read-only API calls to calculate metrics. - Sanitization: The instructions provide specific guidance for URL encoding OData parameters and escaping characters for shell compatibility.
Audit Metadata