| name | np-kb-example-infra-pointer |
| description | TEMPLATE — copy into your content overlay and rename. Points the agent at a piece of infrastructure you own (a site, a server, a self-hosted service) — where the repo is, the stack at a glance, how to reach it, which creds/profile to use. Use when working on that system so the agent doesn't rediscover it each time. |
EXAMPLE: an infrastructure-pointer knowledge skill
Template, not a live skill (see the EXAMPLE design-system skill for how to adopt
it). Copy into $NP_CONTENT_DIR/skills/, rename to np-kb-<system>, fill in your
real details, keep it private (it names your infra). Replace every <…>.
A pointer skill is a few durable facts about a system you own, so the agent starts
from "I know where this lives and how it's built" instead of asking.
at a glance
- Repo:
<github org/repo or path> (<private|public>)
- Stack:
<framework / hosting / key services>
- Reach it:
<URL> / <host:port> / <how to SSH or connect>
- Creds:
<which profile / secret name> — pulled via your secrets workflow, never
inlined here. No raw secrets, tokens, IPs-you-consider-sensitive, or passwords in
a skill.
- Authoritative docs:
<where the real docs live>
When to use
"Working on <system>, deploying it, or wiring something to it" → read this first.
Boundary
This is a pointer (durable orientation), not a runbook. Step-by-step procedures and
anything version-sensitive belong in the repo's own docs or a references/ file.
