skills.rest
Krilliac
KrilliacCOMMUNITY
Explore all skills by this author →
GitHub

sonder-diagnostics-and-tooling

Measure Sonder Runtime health using doctor, preflight, status, logs, and read-only benchmark scripts.

6|3|Updated Jul 3, 2026
Search Tags:#monitoring#logging#observability#health check#diagnostics#ollama#powershell
One-click install
npx skills add https://github.com/Krilliac/Sonder-runtime --skill sonder-diagnostics-and-tooling-krilliac
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sonder-diagnostics-and-tooling
Source: https://github.com/Krilliac/Sonder-runtime/tree/main/.claude/skills/sonder-diagnostics-and-tooling
Command: npx skills add https://github.com/Krilliac/Sonder-runtime --skill sonder-diagnostics-and-tooling-krilliac

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? It replaces guesswork about whether a Sonder Runtime installation is healthy with concrete measurements: consolidated doctor checks, preflight gates, status reports, guarded log inspection, and timestamped JSON health snapshots. ## Core Features & Use Cases - Health instrumentation catalog: Documents the eight doctor checks, preflight gates, smoke test, and status/diagnostics CLI surfaces, including exit-code semantics and what healthy output looks like. - Evidence and log locations: Maps exactly where logs, PID files, debug dumps, and evaluation history land on disk, with redaction guarantees for safe bug-report attachments. - Read-only measurement scripts: Covers benchmark and evaluation scripts plus a bundled PowerShell helper that captures preflight, doctor, and status JSON reports to timestamped files. - Use Case: When a user asks "is the server healthy", run the doctor and status commands, inspect the guarded log tail, and produce a JSON snapshot proving the state of config, storage, schemas, and the Ollama connection. ## Quick Start Run the health snapshot helper to capture preflight, doctor, and status JSON reports into an output folder, then review any FAIL or WARN lines before serving traffic.

Frequently Asked Questions about sonder-diagnostics-and-tooling

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check if the Sonder Runtime server is healthy?▼

Run python -m sonder_runtime doctor from the repo root for an eight-check health rollup covering config, storage, schemas, policy, and Ollama. Exit code is 1 only on fail; warnings exit 0. Add --json for machine-readable output.

How do I capture Sonder health reports to files?▼

Run the bundled health-snapshot.ps1 script with an -OutputDir argument. It executes preflight, doctor, and status with --json and writes each report to timestamped JSON files, with stderr captured separately so diagnostics noise never corrupts the JSON.

What is the difference between doctor, preflight, and smoke?▼

Doctor is a read-only eight-check health rollup that never mutates state. Preflight is the startup gate that creates the state home and runs a write probe. Smoke additionally migrates the operations store and writes one event, proving the storage stack works end to end.

Why does the doctor report a warning when Ollama is reachable?▼

A reachable Ollama endpoint with zero models installed produces a warn result, because the runtime starts but the next chat turn fails at the provider. Install a model with the setup script to clear the warning.

Where do Sonder Runtime logs and evidence files go?▼

Under the state home: run/sonder_serve.log and run/ollama.log hold server output, run/ contains PID files, dumps/ holds redacted debug dumps, and eval-history.jsonl stores append-only evaluation evidence. A loopback dashboard at port 11435 shows a live redacted log tail.

When should I not use this diagnostics skill?▼

Do not use it for symptom-to-cause triage decision trees, which belong to the debugging playbook skill, or for model evaluation and promotion policy, which belongs to the validation and QA skill. This skill only catalogs measurement instruments and evidence locations.

Related Skills

Explore More Software Engineering Skills→
FnSK4R17sFnSK4R17s
1

python-scripts

Execute Python scripts in a temporary virtual environment with dependencies.

Community
Intermediate
kieranpottskieranpotts

fix

Repairs build, lint, and type-check failures with minimal verified changes.

Community
Intermediate
sunholo-datasunholo-data
33

Eval Analyzer

Analyze AILANG evaluation baselines to categorize failures and compare model performance.

Official
Intermediate
HugoAlmeidaMoreiraHugoAlmeidaMoreira

agent-runtime-capability-audit

Diagnose runtime capability gaps in Hermes agent deployments across skills, CLI tools, and network reachability.

Community
Advanced