What problem does it solve? Running raw podman commands forces agents to parse human-shaped output, re-read help text, and guess at next steps. This Skill wraps the Podman CLI and returns structured TOON output with explicit counts, stable key sets, runnable next-step suggestions, and machine-readable error codes an agent can branch on. ## Core Features & Use Cases - Container lifecycle management: list, inspect, run, stop, and remove containers, plus build, pull, tag, and prune images across pods, volumes, networks, and secrets. - Safe mutations: every write subcommand supports --dry-run with a would_* preview block, and secret create only accepts --from-file or --from-env so values never appear in process arguments. - Connection and machine management: resolve connections via flag, environment variables, or local socket, and inspect or start podman machines on macOS and Windows. - Use Case: An agent needs to deploy a test stack. It runs podman-tool container run nginx:latest --name web -p 8080:80 --dry-run to preview the mutation, then executes it and reads the help: block for the exact inspect and logs commands to run next. ## Quick Start Ask the agent to list running containers with podman-tool and follow the suggested next-step commands in the output's help block.
