next-task

This skill's stated purpose (fetching the next task using a local taskmd CLI and reading the resulting task file) aligns with its instructions. The primary security concern is allowing an AI agent to run arbitrary Bash commands and read arbitrary files returned by taskmd. That capability is broad and can lead to local data exposure if taskmd or the provided arguments are manipulated. There are no remote downloads, hardcoded secrets, or obfuscated code in the skill, so the immediate supply-chain risk is low, but the local command-and-file-read pattern elevates the overall risk. Use with caution: restrict arguments, validate file paths, and limit file-read scope when possible.