signals-scout-customer-analytics
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is authored by PostHog and utilizes internal platform tools (
execute-sql,emit_report,scout-scratchpad-remember) to perform its stated purpose of customer analytics. All data access and reporting occur within the authorized PostHog environment. - [DATA_EXFILTRATION]: While the skill accesses sensitive business data such as account properties, engagement volumes, and CRM identifiers (
stripe_customer_id,hubspot_deal_id), this is the primary function of a customer analytics tool. The data is processed internally to generate reports for human CSMs/AEs, with no evidence of exfiltration to unauthorized external domains. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the
system.accountsandeventstables (e.g., account names, user-defined properties). While it lacks explicit boundary markers for this data, the agent's capabilities are restricted to internal reporting and memory tools, significantly limiting the impact of any potential injection via database fields. Severity is considered low and consistent with standard analytical operations. - [REMOTE_CODE_EXECUTION]: No remote code execution patterns, package installations, or unauthorized script downloads were detected. The SQL execution is performed via a read-only analytics tool provided by the platform.
- [OBFUSCATION]: The instructions and code samples are written in clear text with no evidence of Base64, hex encoding, zero-width characters, or hidden acrostics.
Audit Metadata