Replace pipe-delimited runners.conf with INI-style sections supporting
host resolution, container images, repo-scoped tokens, resource limits,
capacity, and SSH key passthrough. All defaults pulled from .env.
- Add INI parsing helpers (ini_list_sections, ini_get, ini_set) to common.sh
- Add SSH key support (UNRAID_SSH_KEY, FEDORA_SSH_KEY) to ssh_exec/scp_to
- Add .env vars: RUNNER_DEFAULT_IMAGE, RUNNER_DEFAULT_CAPACITY,
RUNNER_DEFAULT_DATA_PATH, LOCAL_RUNNER_DATA_PATH, LOCAL_REGISTRY
- Rewrite manage_runner.sh with host/image/token resolution and resource limits
- Rewrite configure_runners.sh wizard for INI format with all 9 fields
- Update phase3 scripts to use ini_list_sections instead of pipe parsing
- Add runners.conf INI validation to preflight.sh (check 5b)
- Update templates to use resolved labels, capacity, and deploy resources
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- .gitignore: excludes .env, runners.conf, certs, temp files, editor files
- .env.example: all configuration variables with sections and descriptions
- runners.conf.example: dynamic runner definition format (pipe-delimited)
- PLAN.md: comprehensive implementation plan with DoD for all 18 milestones
- CLAUDE.md: project conventions and instructions
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
