When using ``-R`` to query a remote transport instance, you must also specify ``-i`` with the path to a management identity file that is authorized for remote management on the target system.
Path found, destination <c89b4da064bf66d280f0e4d8abfd9806> is 4 hops away via <f53a1c4278e0726bb73fcc623d6ce763> on TCPInterface[Testnet/dublin.connect.reticulum.network:4965]
The ``rngit`` utility provides full Git repository hosting and interaction over Reticulum. It allows you to host Git repositories on Reticulum nodes, and to interact with remote repositories using standard Git commands through the ``rns://`` URL scheme.
The system consists of two parts: The ``rngit`` node that hosts repositories, and the ``git-remote-rns`` helper that enables Git to communicate with rngit nodes. As soon as you have RNS installed on your system, you can transparently use Git with Reticulum-hosted repositories just like any other type of remote. Git over Reticulum uses URLs in the following format: ``rns://DESTINATION_HASH/group/repo``.
If you set a branch to track a Reticulum remote as the default upstream, you can simply use `git` as you normally would; all commands work transparently and as expected.
..warning::
**The rngit program is a new addition to RNS!** This functionality was introduced in RNS 1.2.0. While great care has been taken to design a secure, but highly configurable and flexible permission system for allowing many users to interact with many different repositories on a single node, ``rngit`` has not been tested extensively in the wild! Be careful when hosting repositories, especially if they are public or semi-public.
**Usage Examples**
Run ``rngit`` to start a repository node:
..code:: text
$ rngit
[Notice] Starting Reticulum Git Node...
[Notice] Reticulum Git Node listening on <4a5c8d9e1f2a3b4c5d6e7f8a9b0c1d2e>
On the first run, ``rngit`` will create a default configuration file. You will then need to edit this, to point to your repository locations, configure access permissions, and perform any other necessary configuration.
You can run ``rngit`` in service mode with logging to file:
The ``rngit`` node organizes repositories into groups. Each group is a directory containing bare Git repositories. The repository path format is ``group_name/repo_name``. For example, a repository at ``/var/git/public/myrepo`` would be accessible as ``public/myrepo`` via the URL ``rns://DESTINATION_HASH/public/myrepo``.
**Configuration**
The ``rngit`` node configuration file is located at ``~/.rngit/config`` (or ``/etc/rngit/config`` for system-wide installations). The default configuration includes:
- Repository group paths defining where to find bare repositories
- Access permissions for groups and individual repositories
- Announce intervals for network visibility
Access permissions can be configured at the group level in the config file, or per-repository using ``.allowed`` files. Permissions use the format ``permission:target`` where permission is ``r`` (read), ``w`` (write), or ``rw`` (read/write), and target is ``all``, ``none``, or a specific identity hash.
Repository-specific ``.allowed`` files can be static text files or executable scripts that output permission rules to stdout. A ``group.allowed`` file in a repository group directory applies to all repositories within that group.
--config CONFIG path to alternative config directory
--rnsconfig RNSCONFIG
path to alternative Reticulum config directory
-s, --service rngit is running as a service and should log to file
-i, --interactive drop into interactive shell after initialisation
-v, --verbose increase verbosity
-q, --quiet decrease verbosity
--version show program's version number and exit
**All Command-Line Options (git-remote-rns)**
The ``git-remote-rns`` helper is automatically invoked by Git when interacting with ``rns://`` URLs. It is not typically run directly by users, but accepts the following environment variables for configuration:
-``RNGIT_CONFIG`` - Path to alternative client configuration directory
-``RNS_CONFIG`` - Path to alternative Reticulum configuration directory
The client configuration file is located at ``~/.rngit/client_config`` and allows adjusting parameters such as the reference batch size for transfers.
Reticulum includes built-in functionality for discovering connectable interfaces over Reticulum itself. This is particularly useful in situations where you want to do one or more of the following:
* Discover connectable entrypoints available on the Internet
* Find connectable radio access points in the physical world
* Maintain connectivity to RNS instances with unknown or changing IP addresses
Discovered interfaces can be **auto-connected** by Reticulum, which makes it possible to create setups where an arbitrary interface can act simply as a bootstrap connection, that can be torn down again once more suitable interfaces have been discovered and connected.
The interface discovery mechanism uses announces sent over Reticulum itself, and supports both publicly readable interfaces and private, encrypted discovery, that can only be decoded by specified *network identities*. It is also possible to specify which network identities should be considered valid sources for discovered interfaces, so that interfaces published by unknown entities are ignored.
..note::
A *network identity* is a normal Reticulum identity keyset that can be used by
one or more transport nodes to identify them as belonging to the same overall
network. In the context of interface discovery, this makes it easy to manage
connecting to only the particular networks you care about, even if those networks
utilize many individual physical transport node.
This also makes it convenient to auto-connect discovered interfaces only for networks you have some level of trust in.
For information on how to make your interfaces discoverable, see the :ref:`Discoverable Interfaces<interfaces-discoverable>` chapter of this manual. The current section will focus on how to actually *discover and connect to* interfaces available on the network.
In its most basic form, enabling interface discovery is as simple as setting ``discover_interfaces`` to ``true`` in your Reticulum config:
..code:: text
[reticulum]
...
discover_interfaces = yes
...
Once this option is enabled, your RNS instance will start listening for interface discovery announces, and store them for later use or inspection. You can list discovered interfaces with the ``rnstatus`` utility:
Sideband Hub Backbone ✓ Available 1h ago 16 46.2316, 6.0536
RNS Amsterdam Backbone ✓ Available 32m ago 16 52.3865, 4.9037
You can view more detailed information about discovered interfaces, including configuration snippets for pasting directly into your ``[interfaces]`` config, by using the ``rnstatus -D`` option:
In addition to providing local interface discovery information and control, the ``rnstatus`` utility can export discovered interface data in machine-readable JSON format using the ``rnstatus -d --json`` option. This can be useful for exporting the data to external applications such as status pages, access point maps and similar.
To control what sources are considered valid for discovered sources, additional
configuration options can be specified for the interface discovery system.
* The ``interface_discovery_sources`` option is a list of the network or transport identities from which interfaces will be accepted. If this option is set, all others will be ignored. If this option is not set, discovered interfaces will be accepted from any source, but are still subject to stamp value requirements.
* The ``required_discovery_value`` options specifies the minimum stamp value required for the interface announce to be considered valid. To make it computationally difficult to spam the network with a large number of defunct or malicious interfaces, each announced interface requires a valid cryptographical stamp, of configurable difficulty value.
* The ``autoconnect_discovered_interfaces`` value defaults to ``0``, and specifies the maximum number of discovered interfaces that should be auto-connected at any given time. If set to a number greater than ``0``, Reticulum automatically manages discovered interface connections, and will bring discovered interfaces up and down based on availability. You can at any time add discovered interfaces to your configuration manually, to persistently keep them available.
* The ``network_identity`` option specifies the *network identity* for this RNS instance. This identity is used both to sign (and potentially encrypt) *outgoing* interface discovery announces, and to decrypt incoming discovery information.
The configuration snippet below contains an example of setting these additional configuration options:
Reticulum networks are fundamentally permissionless and open, allowing anyone with a compatible interface to participate. While this openness is essential for a resilient and decentralized network, it also exposes the network to potential abuse, such as peers flooding the network with excessive announce broadcasts or other forms of resource exhaustion.
The **Blackhole** system provides tools to help manage this problem. It allows operators and individual users to block specific identities at the Transport layer, preventing them from propagating announces through your node, and for other nodes to reach them through your network.
There is fundamentally **no way** to *globally* block or censor any identity or destination in Reticulum networks. The blackhole functionality will prevent announces from (and traffic to) all destinations associated with the blackholed identity *on your own network segments only*.
This provides users and operators with control over what they want to allow *on their own network segments*, but there is no way to globally censor or remove an identity, as long as *someone* is willing to provide transport for it.
This functionality serves a dual purpose:
***For Individual Users:** It offers a simple way to maintain a quiet and efficient local network by manually blocking spammy or unwanted peers.
***For Network Operators:** It enables the creation of federated, community-wide security standards. By publishing and sharing blackhole lists, operators can protect large infrastructures and distribute spam filtering rules across the mesh without manual intervention.
Local Blackhole Management
==========================
The most immediate way to manage unwanted identities is through manual configuration using the ``rnpath`` utility. This allows you to instantly block or unblock specific identities on your local Transport Instance.
**Blackholing an Identity**
To block an identity, use the ``-B`` (or ``--blackhole``) flag followed by the identity hash. You can optionally specify a duration and a reason, which are useful for logging and future reference.
..code:: text
$ rnpath -B 3a4f8b9c1d2e3f4g5h6i7j8k9l0m1n2o
You can also add a duration (in hours) and a reason:
Manually blocking identities is effective for immediate threats, but maintaining an up-to-date blocklist for a large network is impractical. Reticulum supports **automated list sourcing**, allowing your node to subscribe to blackhole lists maintained by trusted peers, or a central authority you manage yourself.
**Verify Before Subscribing!** Subscribing to a blackhole source is a powerful action that grants that source the ability to dictate who you can communicate with. Before adding a source to your configuration, verify that the maintainer aligns with your usage policy and values. Blindly subscribing to untrusted lists could inadvertently block legitimate peers or essential services.
When enabled, your Transport Instance will periodically (approximately once per hour) connect to configured sources, retrieve their latest blackhole lists, and automatically merge them into your local blocklist. This provides "set-and-forget" protection for both individual users and large networks.
**Configuration**
To enable automated sourcing, add the ``blackhole_sources`` option to the ``[reticulum]`` section of your configuration file. This option accepts a comma-separated list of Transport Identity hashes that you trust to provide valid blackhole lists.
..code:: ini
[reticulum]
...
# Automatically fetch blackhole lists from these trusted sources
1. When enabled, the ``BlackholeUpdater`` service runs in the background.
2. For every identity hash listed in ``blackhole_sources``, it attempts to establish a temporary link to its associated``rnstransport.info.blackhole`` destination.
3. It requests the ``/list`` path, which returns a dictionary of blackholed identities and their associated metadata.
4. The received list is merged with your local ``blackholed_identities`` database.
5. The lists are persisted to disk, ensuring they survive restarts.
..note::
You can verify the external lists you are subscribed to, and their contents, without importing them by using ``rnpath -p``. See the :ref:`rnpath utility documentation<utility-rnpath>` for details on querying remote blackhole lists.
If you are operating a public gateway, a community hub, or simply wish to share your blackhole list with others, you can configure your instance to act as a blackhole list publisher. This allows other nodes to subscribe to *your* definitions of unwanted traffic.
To publish your local blackhole list, enable the ``publish_blackhole`` option in the ``[reticulum]`` section:
..code:: ini
[reticulum]
...
publish_blackhole = yes
...
When this is enabled, your Transport Instance will register a request handler at ``rnstransport.info.blackhole``. Any peer that connects to this destination and requests ``/list`` will receive the complete set of identities currently present in your local blackhole database.
**Federation and Trust**
The blackhole system relies on the trust relationship between the subscriber and the publisher. By subscribing to a source, you are implicitly trusting that source to only block identities that are genuinely detrimental to the network.
As the ecosystem matures, this system is designed to integrate with **Network Identities**. This allows communities to verify that a published blackhole list is actually provided by a specific network or organization with a certain level of reputation and trustworthiness, adding a layer of cryptographic trust to the federation process. This prevents malicious actors from publishing fake lists intended to censor legitimate traffic.
For operators, this creates a scalable model where maintaining a single high-quality blocklist can protect thousands of downstream peers, drastically reducing the administrative.
