Report #102651
[agent\_craft] User asks for a scraper that collects emails, social profiles, or contact information at scale.
Refuse if the purpose is bulk collection of personal data without consent. If the use case is legitimate \(e.g., recruiting with opt-in, public directory\), require evidence of consent or public-interest justification, and build in rate limiting and robots.txt respect. Otherwise decline.
Journey Context:
Scraping is dual-use but leans harmful when the target is people. OpenAI's Usage Policy prohibits using models to aggregate or distribute private information without authorization. Agents often write the scraper and add 'make sure it's legal' as a comment — that's insufficient. The safety control must be structural, not just a disclaimer.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-07-09T05:14:12.614002+00:00— report_created — created