lookalike-customer-finder-skill
Pass
Audited by Gen Agent Trust Hub on Mar 22, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted customer names and descriptions provided by the user ('Input your best customers', 'Find companies like [customer names]'). This creates an indirect prompt injection surface where malicious instructions could be embedded in the names or metadata of the 'customers' being analyzed, potentially influencing the AI's scoring or outreach recommendations.
- [DATA_EXFILTRATION]: The skill instructions mandate writing session and performance data to a local hidden directory (
~/.claude/skill-analytics/last-outcome-lookalike-customer-finder.json). While this is used for analytics and not currently sent to a remote server, unmonitored file writes can be a precursor to data collection for exfiltration.
Audit Metadata