Execute experiments in reproducible Podman containers with provenance capture.
npx skills add https://github.com/gp201/agent-skills --skill podman-runner
Please help me install this Agent Skill. Skill: podman-runner Source: https://github.com/gp201/agent-skills/tree/main/skills/podman-runner Command: npx skills add https://github.com/gp201/agent-skills --skill podman-runner
Run experiments inside reproducible Podman containers to ensure consistent environments and auditable provenance.
Run a sample experiment inside a Podman container to reproduce results and generate a runs/<id>/ provenance directory.
High-intent search queries and answers about installing and using this skill.
Run reproducible experiments by executing commands inside Podman containers that generate a unique run-id and capture provenance artifacts like config.yaml, git.txt, env.txt, and logs within a runs/<id>/ directory. This ensures consistent environments and easy reproduction of prior runs.
Provenance data captured includes config.yaml, git.txt, env.txt, cmd.txt, and comprehensive execution logs. These artifacts are stored within the runs/<id>/ directory to provide an auditable record of the environment, git status, and command invocation for each experiment.
Yes, GPU passthrough is supported when available during containerized experiment execution. The runner automatically coordinates GPU access during the execution phase to enable hardware-accelerated workloads while maintaining reproducible provenance capture.
Yes, it builds or uses a Containerfile-based image to execute commands in a reproducible workspace. The Skill coordinates image building to ensure the environment remains consistent across multiple experiment runs.
Reproduce prior runs by utilizing the provenance artifacts stored in the runs/<id>/ directory, which contains the exact config, git status, environment variables, and command invocation. This enforces a reproducible workspace by binding the project directory for containerized execution.
observerwCreate and migrate AI skills using Containerfile and OCI images.
WangJie-cnTrack ML experiment configurations, metrics, and GPU usage.
jonnymuirScan staged files for credentials and block commits containing secrets.
drpedapatiRecord experiment provenance and reproducible run metadata for research workflows.
