understanding-billing-usage
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes data that may contain user-generated content, such as event names, property values, and log entries, creating a potential surface for indirect prompt injection.
- Ingestion points: Data is pulled from project-specific event tables and logs via SQL and tool calls in
SKILL.md. - Boundary markers: The instructions in
SKILL.md(Step 5) explicitly mandate that the agent treat this data as 'untrusted evidence' and ignore any instructions or guidance embedded within it. - Capability inventory: The skill utilizes MCP tools for billing overview, usage, and spend, along with various product-specific analysis tools for events, replays, and logs.
- Sanitization: The skill employs SQL logic to filter out specific non-billable system events, narrowing the scope of ingested data to relevant metrics.
Audit Metadata