# Bus Audit

> Run OmniClaude bus health audit (Layer 2 domain validation)

- Skill: `omninode-ai/bus-audit` (Agent Skill)
- Install (CLI): `npx skillmds@latest add omninode-ai/bus-audit`
- Raw SKILL.md: https://api.skillmd.com/api/skills/omninode-ai/bus-audit/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: OmniNode-ai (https://skillmd.com/u/omninode-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/omninode-ai/bus-audit

---


<!-- routing-enforced: dispatches to node_bus_audit_compute. Fully implemented. -->

# Bus Audit

**Usage:** `/bus-audit [flags]`

Dispatch to `node_bus_audit_compute` in omnimarket. Do not reimplement bus audit logic inline.

## Announce

Say: "I'm using the bus-audit skill to dispatch node_bus_audit_compute."

## Dispatch

```bash
onex run node_bus_audit_compute -- \
  ${JSON:+--json} \
  ${FAILURES_ONLY:+--failures-only} \
  ${VERBOSE:+--verbose} \
  ${SKIP_DAEMON:+--skip-daemon} \
  ${BROKER:+--broker "$BROKER"} \
  ${SAMPLE_COUNT:+--sample-count "$SAMPLE_COUNT"}
```

Surface the JSON output from stdout. The node produces a `ModelSkillResult` with `status`, `run_id`, and `message`.

On non-zero exit, a `SkillRoutingError` JSON envelope is returned — surface it directly, do not produce prose. If dispatch cannot execute, report the error and stop.

Never re-implement bus audit logic inline. If the node is unavailable, stop — do not fall back to inline script execution or prose audit.