Specter I — OSINT

Most OSINT training stops at typing queries into search engines — ten percent of the job. The other ninety — source independence, calibrated confidence, OPSEC against a target who counter-investigates, defensible documentation — is where professionals live and where almost no public course goes.

Specter I goes there. By the capstone, your findings hold under legal scrutiny, your tradecraft survives adversarial counter-intelligence, and your report meets Berkeley Protocol standards.

• Operational discipline graded throughout. EXIF leaks, persona-real-account cross-pollination, query timing patterns — all detected and scored. No other training treats analyst OPSEC as a graded outcome.
• Adversarial targets. From level twelve onward you face counter-intelligence: canary tokens, dangle accounts, watermarked documents, fake leaks. Detect the trap or trip the alarm.
• Calibrated confidence required. Every claim from level four onward carries a word-of-estimative- probability rating (Admiralty A1–F6). Overconfident wrong answers cost more than honest uncertainty.
• Berkeley Protocol report at graduation. The capstone requires a written intelligence package with chain of custody, source documentation, and alternative-hypothesis consideration. Defensible methodology, not just a flag.
• Quarterly errata, public dashboard. OSINT tools rot fast. Twitter API, CrowdTangle, half of 2018's stack — all dead. Our errata page is public; we re-shoot levels when the underlying primitive shifts. No 2018 advice in 2026 wrapping.