The Agent Skills Directory

[DYNAMIC_EXECUTION]: The rngit node can be configured to execute local files to resolve permissions and generate content.
Executable .allowed files: If a permission file is made executable, the node runs it and parses its standard output as the permission rules for a group or repository.
Executable templates: In the Nomad Network page node configuration, executable templates in the ~/.rngit/templates/ directory are run, and their standard output is used as the template content.
[COMMAND_EXECUTION]: The skill provides instructions to modify global Git configurations to use rngcs as a custom signing program.
Instructs setting gpg.ssh.program to rngcs, which causes Git to execute this shim during every commit signing and verification operation.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install external Python packages to enable functionality.
Fetches the rngit package using pip install rngit or pipx install rngit.
Fetches the pygments package for syntax highlighting on Nomad Network pages using pip install pygments.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data that could potentially contain malicious instructions intended to influence the agent or the rendering node.
Ingestion points: Untrusted data enters the context through Reticulum-hosted Git repositories, work documents (msgpack format), and releases containing Markdown/Micron files.
Boundary markers: The system uses Ed25519 cryptographic signatures (RSG) to verify the integrity and attribution of work documents and releases, but does not specify delimiters for natural language instructions inside these documents.
Capability inventory: The environment includes command execution via the rngit and rngcs CLI tools, as well as the execution of local scripts for permissions and templates.
Sanitization: The skill describes automatic Markdown to Micron conversion for rendering pages over the Nomad Network.

rngit