← ClaudeAtlas

nodus-troubleshootinglisted

Use when nodus add, sync, doctor, update, or relay is failing and you need a structured debugging workflow based on workspace state, lock state, package discovery, and managed outputs.
nodus-rs/nodus · ★ 5 · AI & Automation · score 75
Install: claude install-skill nodus-rs/nodus
# Nodus Troubleshooting Debug Nodus failures from first principles: workspace state, dependency resolution, lock state, discovery results, managed outputs, and relay state. ## Workflow 1. Identify the failing stage. - Parse or validation error - Dependency resolution error - Sync collision or managed file drift - Relay state or local checkout mismatch 2. Inspect the package or workspace state. - Read `nodus.toml` and `nodus.lock`. - Run `nodus info <alias-or-path>` for the failing package. - Run `nodus doctor` for workspace verification. 3. Check for the common root causes. - invalid dependency source or ref - unsupported or duplicate component selection - invalid `managed` mappings - duplicate artifact ids across discovery roots - pending relay edits blocking sync - lockfile drift when using `--locked` or `--frozen` 4. Repair with the smallest correct action. - rerun `nodus sync` - rerun `nodus update` - inspect with `nodus info <alias-or-path>` - refresh the manifest entry - fix the package layout - remove or adopt conflicting managed paths only when the ownership is clear - only if normal recovery fails, consider a last-resort reset: delete `nodus.lock`, clear Nodus-managed outputs, then reinstall or resync 5. Verify again with `nodus doctor`. ## Debug Commands ```bash nodus info . nodus info <alias> nodus doctor nodus sync nodus sync --locked nodus sync --frozen nodus outdated nodus update ``` ## Decision R