kpi-dashboard
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface through the ingestion of external data from user-specified websites, CSV exports, and call transcriptions. However, the accompanying security reference (
references/agent-security.md) provides specific and robust mitigations, such as treating all external content strictly as data, escaping all interpolated values in emitted markup, and explicitly ignoring any instructions found within fetched data. - [DATA_EXFILTRATION]: The skill accesses network resources to perform business research and fetch metrics via the Intempt MCP (Model Context Protocol). These actions are limited to the skill's primary function of dashboard design. Security rules in the provided references prohibit the persistence of credentials and require the redaction of tokens from all outputs.
- [PROMPT_INJECTION]: Static detectors flagged instructions regarding the suppression of warnings. Analysis confirms these are benign process requirements focused on quality assurance, instructing the agent to resolve missing information rather than providing caveated, incomplete drafts to the user, which prevents user reliance on unverified data.
Audit Metadata