Write official Nx AI configuration into the workspace with the official command

Who it's for

You already have an Nx workspace and want to add official Nx configuration artifacts to the AI coding assistants you already use. Use this skill. The source describes it as official AI agent configuration artifacts: skills, subagents, and commands that enhance coding assistants with Nx-specific capabilities. Official repository:

https://github.com/nrwl/nx-ai-agents-config

Preconditions: the current directory is already an Nx workspace, and you accept that running the official command will write the matching configuration files. If you have no workspace, or you are not ready to change local configuration, do not continue yet.

The source mentioned support for multiple agents, and also CI monitoring and self-heal. Do not copy product names that appeared into the body as an exclusive scope, and do not expand "Nx-specific capabilities" into concrete operations that the excerpt never showed. The license is MIT, but that does not replace confirming a write to disk. If you only want to read the repository first and are not ready to land configuration in the workspace, open the official page; you need not run the command immediately. If your goal is to let an agent watch CI and apply patches automatically, remember first that the source writes self-heal as automatically applying fixes. The first try must not take that path.

First try in 3 minutes

First confirm you are inside an Nx workspace, not an ordinary folder. Open the official repository and check:

https://github.com/nrwl/nx-ai-agents-config

Recommended install is only this one. It writes configuration files for your preferred AI tools; before you run it, confirm you want those files left in this workspace:

nx configure-ai-agents

Do not invent a second installer, and do not add a skills add or directory that never appeared. The source does not write concrete paths for the configuration files; do not invent a landing location.

The first time in the agent, only check; do not start CI monitoring, and do not let it auto-fix failures. Ask the agent to state: these are official configuration artifacts, the command will write configuration files for your preferred tools, and capability is written only as "Nx-specific capabilities". The observation is that it can recite those boundaries, has not changed CI, and has not auto-committed fixes.

Success signal: configure-ai-agents ends normally, or the agent can accurately recite that matching configuration files will be written. The source gives no command switches; do not invent parameters. The first time, do not demonstrate self-heal.

Use cases

Case 1: Only write official configuration in an existing workspace. After confirming the current directory is an Nx workspace and you accept a write to disk, run nx configure-ai-agents. Fits people who only want official skills, subagents, and commands attached to an existing assistant and are not ready to touch CI. Success signal: the command ends normally. The source does not write which filenames it will generate; do not invent them. If the command fails, stop; do not switch to an install path that never appeared.

Case 2: First check whether self-heal would change your code. After the configuration is in place, have the agent recite CI monitoring and self-heal from the source: watch CI pipelines, find failures, and automatically apply fixes. Fits people who worry the agent will change the repository immediately. Success signal: it can say this is automatic repair, and this time it did not actually watch or patch. The source gives no command to start or stop self-heal; do not invent one. Do not expand "Nx-specific capabilities" into concrete operations the excerpt never showed.

Boundaries

The main path is only nx configure-ai-agents, and it must run inside an Nx workspace. It writes configuration files for your preferred AI tools. The source does not write those file paths or how to undo them; do not add uninstall steps.

The source mentioned support for multiple agents, but that is also a roster: do not copy it into the body as the only usable scope, and do not apply the configure command to a directory that is not an Nx workspace. CI monitoring and self-heal watch CI pipelines, detect failures, and automatically apply fixes. Side effects are large; do not enable them the first time. Evidence also gives no enable method.

Capability is written only as official configuration artifacts, skills, subagents, commands, and Nx-specific capabilities. Do not fill in unwritten concrete operations or directories. The license is MIT. First-time success is only the configure command finishing, or the agent reciting write-to-disk and self-heal boundaries; before truly auto-fixing CI you must confirm separately. The source does not write how to choose preferred tools; the command writes the matching configuration files itself. Do not invent interactive steps. If the current directory is not an Nx workspace, stop first; do not take this command to an ordinary repository. The first time, do not run in parallel an installer that never appeared. Evidence stops at the words "Nx-specific capabilities" and did not expand an inventory; do not write unwritten operations as already executable.