cli-for-agents

cli-for-agents is an agent skill from Cursor's plugins collection for designing or reviewing command-line tools so that coding agents can run them reliably. Human-oriented CLIs often block agents: they stop for interactive prompts, rely on large upfront documentation, or show help text without examples that can be copied. The skill's patterns are non-interactive first, so every input can be given as a flag or flag value, with interactive prompts only as a fallback when flags are missing. It asks for discoverability without dumping context: an agent discovers subcommands step by step, from the tool name to the help for one subcommand, so each subcommand owns its documentation and unused commands stay out of the context. Help should work and include copy-pasteable examples in layers. The skill also covers stdin and pipelines so the tool composes with other tools, and fast, actionable errors. Its examples contrast a bad prompt-driven deploy command with a good one that takes an environment flag. It is useful for teams building internal tools that agents will run in scripts or continuous integration. It is listed on skills.sh, where the page shows install counts and GitHub stars for the Cursor plugins repository, and its security audits all pass. Install it with npx skills add https://github.com/cursor/plugins --skill cli-for-agents.…