agentleFS
ss2codess2code/nanoclaw-ops/.claude/skills/nano-pvt-hub/SKILL.md

nano-pvt-hub

Skill1 starsChanged 43 days agoMITView on GitHub

Publish to and browse the NanoClaw private docs hub — a tailnet-only document repository (NanoClaw Docs, Dashboards, Trackers, Agent Docs) served under the Ops Center Docs tab. Use to publish/update/list/sweep hub documents from the host, or to (re)scaffold the hub store. Triggers on "docs hub", "nano-pvt-hub", "publish a dashboard/tracker/doc", "private hub".

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every published artifact needs a kind, a title, and a bounded source file.

    line 29
  2. Very specific:

    Always pass --root data/hub.

    line 36

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

nano-pvt-hub — private docs hub (host / clot side)

A single-root document repository browsable over the tailnet from the Ops Center Docs tab (http://127.0.0.1:10333/hub/). It holds four surfaces, all under data/hub/:

SurfaceWhereWritten by
NanoClaw Docsdata/hub/nanoclaw-docs → symlink to repo docs/nobody — read-only index (don't regenerate docs)
Dashboardsdata/hub/dashboards/<audience>/<id>/agents & host (HTML the agent builds for a user)
Trackersdata/hub/trackers/<audience>/<id>/agents & host (status page / append-log / table)
Agent Docsdata/hub/agent-docs/<audience>/<id>/agents & host (generic reference docs)

The engine is one dependency-free script, container/skills/nano-pvt-hub/hub.mjs (the same file agents run inside containers as bun /app/skills/nano-pvt-hub/hub.mjs). .claude/skills/nano-pvt-hub/hub.mjs is a symlink to it — one source of truth.

Manifests are truth; catalog.json is a derived cache (rebuilt atomically on every write). Ops Center reads catalog.json to render the Docs tab; it never imports this script.

Publishing rules (hygiene contract — mirrors hermes-private-hub)

Every published artifact needs a kind, a title, and a bounded source file. Never publish credentials, API keys, raw channel identifiers, chat/runtime DBs, private portfolio data, PII, or raw provider responses. Documents are durable by default (no expiry); pass --ttl 30d to make one expire.

Usage (host)

Always pass --root data/hub. Run from the repo root.

# Publish a dashboard (HTML) or an agent doc (.md/.html/.txt/.json/.csv/.svg)
node container/skills/nano-pvt-hub/hub.mjs publish \
  --kind dashboard --title "Fleet Status" --summary "live fleet view" \
  --source /path/to/report.html --root data/hub

# Update a status page in place — reuse the same --id
node container/skills/nano-pvt-hub/hub.mjs publish \
  --kind tracker --shape status --id fleet-status --title "Fleet Status" \
  --source /path/to/report.html --root data/hub

# Append to a log tracker (auto-creates it on first append)
node container/skills/nano-pvt-hub/hub.mjs append \
  --id deploy-log --title "Deploy Log" --record '{"event":"deploy","ok":true}' --root data/hub

# List / inspect / clean up
node container/skills/nano-pvt-hub/hub.mjs list --root data/hub
node container/skills/nano-pvt-hub/hub.mjs get --kind dashboard --audience shared --id fleet-status --root data/hub
node container/skills/nano-pvt-hub/hub.mjs sweep --root data/hub    # remove expired
node container/skills/nano-pvt-hub/hub.mjs help

publish also accepts --audience <name> (default shared), --series, --ttl none|30d. Trackers accept --shape status|log|table.

Scaffold / apply (idempotent — safe to re-run)

data/hub/ is runtime state (gitignored), so recreate it on any fresh checkout. This creates the store, the four surfaces, the read-only docs symlink, and the initial catalog:

mkdir -p data/hub/dashboards data/hub/trackers data/hub/agent-docs
printf 'NanoClaw nano-pvt-hub artifact store\n' > data/hub/.hub-store
# nanoclaw-docs points at the repo docs/ so all four surfaces sit as siblings
[ -e data/hub/nanoclaw-docs ] || ln -s ../../docs data/hub/nanoclaw-docs
node container/skills/nano-pvt-hub/hub.mjs rebuild --root data/hub

Then restart Ops Center so it serves the new tab (label is discovered from the running service, so this works on any install):

# macOS (launchd)
launchctl kickstart -k gui/$(id -u)/$(launchctl list | grep -o 'com.nanoclaw.opscenter-[a-z0-9]*' | head -1)
# Linux (systemd): the unit is nanoclaw-opscenter-<slug> — find it with:
#   systemctl --user list-units 'nanoclaw-opscenter-*' --no-legend | awk '{print $1}' | xargs -r systemctl --user restart

Browse it at http://127.0.0.1:10333/hub/. For off-machine access over the tailnet, expose the whole Ops Center once (Funnel stays OFF):

tailscale serve --bg --https=443 http://127.0.0.1:10333
tailscale serve status     # expect :443 -> 127.0.0.1:10333
tailscale funnel status    # must be empty

Then browse https://<magicdns-name>/hub/ from any tailnet device.

Notes

  • Agents publish the same way, from inside their container, using bun /app/skills/nano-pvt-hub/hub.mjs ... --root /workspace/hub. See the container skill of the same name.
  • Retrieval/indexing for agents is hub.mjs index (JSON: catalog + docs file list) — a filesystem read, no HTTP needed.
  • Removing the feature: see REMOVE.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.