package token_cache import ( "github.com/MakeNowJust/heredoc/v2" "github.com/spf13/cobra" "gitlab.com/gitlab-org/cli/internal/cmdutils" tokenCacheClearCmd "gitlab.com/gitlab-org/cli/internal/commands/cluster/agent/token_cache/clear" tokenCacheListCmd "gitlab.com/gitlab-org/cli/internal/commands/cluster/agent/token_cache/list" ) func NewCmd(f cmdutils.Factory) *cobra.Command { cmd := &cobra.Command{ Use: "token-cache [flags]", Short: `Manage cached GitLab Agent tokens.`, Long: heredoc.Docf(` Tokens created by %[1]sglab cluster agent get-token%[1]s are cached locally in the system keyring and filesystem to avoid creating new tokens for each kubectl operation. The GitLab CLI caches agent tokens in two locations: - Keyring: Uses the system keyring (Windows Credential Manager, macOS Keychain, Linux Secret Service). - Filesystem: Stores tokens in the user's cache directory as encrypted files. The cache improves performance by avoiding the need to create new tokens for each kubectl operation when using %[1]sglab cluster agent update-kubeconfig%[1]s. Cached tokens are stored using a key format that includes: - Base64-encoded GitLab instance URL - Agent ID This ensures tokens are properly isolated by GitLab instance and agent. `, "`"), } cmd.AddCommand(tokenCacheListCmd.NewCmd(f)) cmd.AddCommand(tokenCacheClearCmd.NewCmd(f)) return cmd }