execute-tasks

SUSPICIOUS: The skill is broadly coherent for autonomous repo task execution, but its autonomy is high: it can modify code, run project commands, and create commits based on local planning files. Main risks are repo-script execution via npx/npm, instruction-following from mutable local markdown/text files, and optional reliance on a secondary Playwright skill. No clear credential harvesting, exfiltration, or malicious data routing is present.