auto-research-public

SUSPICIOUS: the skill is internally coherent for cold-email automation, and its credential/API scope is mostly proportionate, but it enables high-impact autonomous outbound messaging by creating and activating live Smartlead campaigns and encourages recurring unattended runs. Main concern is autonomy and operational abuse risk, not confirmed malware or covert exfiltration.

No strong evidence of intentional malware/backdoor behavior in this module (no obfuscation, dynamic execution, or persistence). However, it presents meaningful security and privacy risks: (1) it fetches https://{company_domain} using unvalidated input, creating an SSRF-style risk if the input JSON can be attacker-controlled; and (2) it transmits lead PII (names/domains/emails) to third-party services and writes the full lead dataset (allLeads) to disk, increasing data-retention exposure. These are primary reasons this should be reviewed in the context of input trust boundaries and data-handling requirements.