A benchmark harness for measuring how AI coding agents find and fix security vulnerabilities.
12 files for Claude Code, Codex and OpenCode: benchmark-add-new-category, benchmark-add-new-fixture, benchmark-chart-generator, benchmark-fixture-statistics and 8 more — 3,693 tokens loaded in every session.
AGENTS.md A 3,690 tok CLAUDE.md A 3 tok .agents/skills/benchmark-add-new-category/SKILL.md A 147 tok .agents/skills/benchmark-add-new-fixture/SKILL.md A 192 tok .agents/skills/benchmark-chart-generator/SKILL.md B 181 tok .agents/skills/benchmark-fixture-statistics/SKILL.md A 96 tok .agents/skills/benchmark-latex-writer/SKILL.md B 89 tok .agents/skills/benchmark-report-writer/SKILL.md A 184 tok .agents/skills/benchmark-run/SKILL.md A 189 tok .agents/skills/blog-writer/SKILL.md A — .agents/skills/cloc/SKILL.md A 80 tok .agents/skills/skill-creator-extra/SKILL.md A 126 tok These files are snyk-labs/snyk-vulnbench's own configuration — they tell Claude Code, Codex and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.