← Capabilities
LLM safety and red-teaming
not yet earnedI can red-team a system I built, handle personal data deliberately rather than incidentally, and show the attacks that still work.
What this claim costs
- 2 assessed items at tier 2 or above0 of 2 — 2 more to do, from 7 available
- A public artifact built for this claim — red-team-reportNot built yet
- A documented negative result — where this stops workingNothing recorded. A capability with no known failure boundary has not been explored.
- Defended aloud, against questions you did not write2 available, none done
- Recalled cold, a week later, without notes1 scheduled, none done
The proof
red-team-report
An adversarial pass over a live product, with the successful attacks published alongside the mitigations.
What it has to contain
- An adversarial pass over a live system, with every input path enumerated
- Successful attacks published verbatim, not only the mitigated ones
- Personal data handling decided per category, including logs
Where it stops working. Not established yet. Anyone can ship a demo; the boundary is the part that is worth reading, so this claim is not earned until it is written down.
Skills underneath
- Alignment techniques
- Red-teaming and adversarial testing
- Privacy and data handling
7 assessed items across the specified courses evidence these, 0 done.