
Your best engineer couldn't pass your own interview
Algorithm puzzles test memorization, and AI can solve them anyway. SkillFoundry gives candidates a real task in a real open-source codebase, with an AI assistant at their side, and measures how they work. Sometimes that assistant is wrong on purpose: we plant a subtle flaw in a suggestion and tell you whether the candidate caught it or shipped it. You get a hiring report with timestamped evidence, not a pass/fail score.
SkillFoundry provides a platform for evaluating developer candidates through real tasks in open-source codebases, using an AI assistant that includes intentional flaws to assess problem-solving skills. The hiring report generated includes timestamped evidence of the candidate's performance rather than a simple pass/fail score.
A source that found nothing is a measurement. A source that has not run is a gap. Neither means the launch lacks the thing.