ff-oce-dashboard
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Ingestion of External Data (Indirect Prompt Injection): The skill processes data from various external platforms, which is a common pattern for dashboard generation but introduces a potential surface for indirect instructions.
- Ingestion points: Data is retrieved from IcM incidents, Azure DevOps (ADO) pipeline results, Kusto query outputs, Teams channel messages/replies, and WorkIQ summaries via Step 1 and Step 2.
- Boundary markers: The instructions do not explicitly define delimiters or provide specific 'ignore' instructions for the agent when processing the text content returned from these services (such as incident titles or Teams message bodies).
- Capability inventory: The skill utilizes MCP tools to query remote APIs and provides an option in Step 0 to write the final dashboard to a local markdown file (e.g.,
oce-dashboard-*.md) based on user choice. - Sanitization: There are no specified sanitization or validation steps for the content ingested from the external services before it is consolidated into the final report.
- Access to Internal Services: The skill queries several organizational endpoints, including Microsoft Kusto and Azure DevOps. These operations are performed using standard tool calls to gather operational data for the dashboard.
Audit Metadata