Use --json for machine-readable results and --yes for destructive
operations in non-interactive contexts.
Automation should be explicit about profile, repository, output mode, and confirmation. Avoid relying on the active local profile or current working directory unless the script is intentionally repository-local.
GITFLEET_GITHUB_TOKEN=github_pat_... gitfleet repo list --json
GITFLEET_PROFILE=work gitfleet issue list --repo owner/repository --limit 50 --page 1 --json
gitfleet repo clone --all --org platform --directory repos --dry-run --json
gitfleet repo delete owner/old-repository --json --yesStatus and debug logs go to stderr. Structured output goes to stdout.
For commands that expose both --limit and --page, treat --limit as the
page size and --page as the 1-based provider page to request. Use explicit
pages in scheduled jobs so each run consumes a predictable slice of issues,
changes, releases, deployments, projects, packages, discussions, pipelines, or
search results.
Use --dry-run before supported bulk mutations:
gitfleet workspace archive platform --dry-run --jsonUse --dry-run before bulk local operations too:
gitfleet repo clone --all --user alice --directory user-repos --dry-run --jsonParse stdout for JSON results and leave stderr for progress or debug logs. If a
job needs diagnostics, add --debug and keep logs redacted before sharing them.