pipeline-yield-analytics-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from external sources (HubSpot deals and Clari call records), creating a surface for indirect prompt injection attacks.\n
- Ingestion points: Data is pulled into the agent context via MCP tools including
hubspot_search_deals,clari_list_calls,clari_get_call_summary, andquery_dataset.\n - Boundary markers: The instructions do not specify the use of delimiters or provide instructions to the model to treat the ingested text as untrusted or separate from the skill's primary objectives.\n
- Capability inventory: The skill possesses the capability to write Excel workbooks using the
xlsxskill, create live dashboards viacreate_artifact, and access sensitive financial data viarevenueanddeals_closeddatasets.\n - Sanitization: While the skill uses
data:validate-datato verify the mathematical accuracy of the funnel metrics, it lacks specific sanitization or filtering for the textual content retrieved from CRM records that might contain adversarial prompts.
Audit Metadata