flaky-smoke-tests

A guide for investigating intermittent failures in VS Code’s Electron smoke-test pipeline. A smoke test is a broad check that a built application starts and its main functions work.

In plain words
What is it for?
It supports finding failed iterations, downloading Azure DevOps logs and platform artifacts, comparing runner logs, tracing the introducing change, and starting focused follow-up tests.
Why use it?
It prevents the overall pipeline result from hiding failed attempts and helps separate logs from multiple test runs.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/microsoft/vscode/flaky-smoke-tests
Any agent
npx skills add microsoft/vscode --skill flaky-smoke-tests
Clone the repo
git clone --depth 1 https://github.com/microsoft/vscode

Made for: Claude Code, Codex.

Per session 63 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,067 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00063 $0.03067
Opus 5 $0.00032 $0.01533
Sonnet 5 $0.00013 $0.00613
Haiku 4.5 $0.00006 $0.00307

Measured 3d ago against content hash 5e1a305d4585, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

flaky-smoke-tests scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.github/skills/flaky-smoke-tests/SKILL.md · 352 lines

How it starts

The opening of the file, as written. The whole thing — 352 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Diagnosing Flaky Smoke Tests

Use this skill for failures from the Azure DevOps Flaky Smoke Tests pipeline:

  • Organization: https://dev.azure.com/monacotools
  • Project: Monaco
  • Definition ID: 700
  • Pipeline YAML: build/azure-pipelines/product-smoke-flaky.yml
  • Scheduled runs: twice daily on main

The pipeline builds VS Code from source and runs the complete Electron smoke suite once per entry in the iterations parameter. Each platform runs its iterations sequentially in one job.

Important Pipeline Behavior

  • Each iteration is a separate timeline task named Smoke test iteration <N>/<total> (Electron).
  • Iteration tasks use continueOnError: true, so a failed iteration normally has result succeededWithIssues and later iterations still run.
  • Do not infer smoke-test health from the overall build or job result. Inspect every iteration task.
  • Logs are published once per platform job from the shared .build/logs directory. The artifact is job-scoped, not iteration-scoped: smoke-test-runner.log can contain multiple iterations, and suite directories are not separated by iteration.
  • Do not cancel a platform job while investigating if you still need its logs artifact. Cancellation can skip the publish steps entirely, even when one or more failed iteration task logs are already available. If a validation run has become predictably red, either let the current job finish or queue a single diagnostic iteration and let that job publish before iterating.
  • Later iterations can overwrite or truncate files in the shared logs directory. The final artifact may therefore contain the last iteration's suite directory or runner log but not the failed iteration's detailed diagnostics. Preserve the exact task log, and use a one-iteration diagnostic run when extension-host, renderer, mock-server, screenshot, or trace evidence from the failed attempt is required.
  • The exact iteration task log is authoritative for its failure summary and time range. Correlate that range with the job-level artifact.
  • Iterations can see persisted smoke-test state from earlier iterations, such as historical session rows. Never assume the newest-looking row or a matching prompt belongs to the active conversation without verifying the active view.

Read the full file on GitHub · 352 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 352 lines · 63 tokens per session scan A 5e1a305d4585

Subscribe to this mod's changes

flaky-smoke-tests is a skill published in the GitHub repository microsoft/vscode (190,384 stars, last pushed yesterday), licensed MIT. It adds 63 tokens to every session and 3,067 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories