VideoVideo:AI Native Dev
Benchmarking and Refining a Snyk AI Skill Against Anthropic Rules
Simon Maple (Tessl) and Krzysztof Huszcza (Snyk) evaluate and optimize an existing Snyk AI skill repository. Using the Tessl agent, they grade the skill against Anthropic's guidelines, automate refactoring to avoid context bloat via progressive disclosure, and discuss skill security with Snyk's Agent Scan.