Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Cloud Development Environment · elie222 | DecimalAI

elie222/cloud-dev-environment

SkillScore58Maintained

Cursor Cloud VM setup and service startup instructions for local development

DevOpsby elie222

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

DevOps & SRE pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/elie222-cloud-dev-environment/SKILL.md

decimalai skills pull elie222-cloud-dev-environment --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/elie222-cloud-dev-environment/SKILL.md https://app.decimal.ai/s/elie222-cloud-dev-environment/SKILL.md

Raw file: /s/elie222-cloud-dev-environment/SKILL.md ↗ · machine-readable summary at /s/elie222-cloud-dev-environment.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

24 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 30d ago

—-11% tokens0% turns58% of whole cases pass with skill (14/24)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-01	✗→✓	▲ Improved	68%	0%
case-04	✗→✓	▲ Improved	-36%	0%
case-06	✗→✓	▲ Improved	-22%	0%
case-07	✗→✓	▲ Improved	-11%	0%
case-08	✗→✓	▲ Improved	10%	0%

Showing 5 of 24 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

elie222 / inbox-zero

.claude/skills/cloud-dev-environment/SKILL.md

View on GitHub ↗

License NOASSERTIONImported 8/10/2026Version v1

Skill contents

SKILL.md

Cloud Development Environment

Services overview

Main app (apps/web): Next.js 16 app (Turbopack). Runs on port 3000.
PostgreSQL 16: Primary database. Runs on port 5432 via docker-compose.dev.yml.
Redis 7 + serverless-redis-http: Caching/rate-limiting. Redis on port 6380, HTTP proxy on port 8079.

Starting services

Start Docker daemon if needed: sudo dockerd &>/dev/null &, then confirm with docker info (or sudo docker info if the client cannot reach the daemon yet).
Start databases from repo root: docker compose -f docker-compose.dev.yml up -d. If you get permission denied on /var/run/docker.sock, run the same command with sudo docker compose instead of making the socket world-writable.
Run Prisma migrations: cd apps/web && pnpm prisma:migrate:local (uses dotenv -e .env.local; do NOT use bare prisma migrate dev — it won't load .env.local).
Start dev server: pnpm dev from repo root.

Environment file

The app reads apps/web/.env.local. Required non-obvious env vars beyond .env.example defaults:

DEFAULT_LLMS (e.g. openai:gpt-5.4-mini) — app crashes at startup without this.
GOOGLE_CLIENT_ID and GOOGLE_CLIENT_SECRET — hard-required by env.ts validation. Use Google emulator credentials for local dev (see below).
UPSTASH_REDIS_TOKEN must be set to dev_token to match the SRH_TOKEN default in docker-compose.dev.yml.
MICROSOFT_WEBHOOK_CLIENT_STATE — required if MICROSOFT_CLIENT_ID is set.

Google emulator for local dev

Use the Google emulator instead of real OAuth credentials:

GOOGLE_CLIENT_ID=emulate-google-client.apps.googleusercontent.com
GOOGLE_CLIENT_SECRET=emulate-google-secret
GOOGLE_BASE_URL=http://localhost:4002

Start the emulator: docker compose -f docker-compose.dev.yml --profile google-emulator up -d.

Testing

pnpm test runs Vitest unit/integration tests (no DB or external services required).
pnpm lint runs Biome. Pre-existing lint warnings/errors in the repo are expected.
AI tests (pnpm test-ai) require a real LLM API key and are skipped by default.

Docker in this environment

The cloud VM is a Docker-in-Docker setup. Docker requires fuse-overlayfs storage driver and iptables-legacy. These are configured during initial setup. After snapshot restore, run sudo dockerd &>/dev/null & if Docker daemon is not running.

The VM may not have the Docker Compose v2 plugin pre-installed. If docker compose version fails, install it:

bash
sudo mkdir -p /usr/local/lib/docker/cli-plugins
sudo curl -fsSL https://github.com/docker/compose/releases/download/v2.29.2/docker-compose-linux-x86_64 -o /usr/local/lib/docker/cli-plugins/docker-compose
sudo chmod +x /usr/local/lib/docker/cli-plugins/docker-compose

Onboarding flow

After a fresh login via the Google emulator, the app forces an onboarding wizard. Some onboarding buttons require JavaScript click() calls rather than standard browser clicks (React event delegation quirk in headless/automation contexts).

Illustrative Want to see what elie222/cloud-dev-environment would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/29/2026, 04:13 UTC · 24 cases

Measured on 20 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

+29%

58% vs 29%

Agent turns

0%

1.0 vs 1.0

Tokens

-11%

1,593.55 vs 1,793.8

fail → pass · 9pass → pass · 5fail → fail · 8pass → fail · 2

Per-case results (24)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I just logged into a fresh cloud development VM and need to get the whole local stack up and running. Could you provide a clear execution plan listing the exact shell commands in order to bring up the background services, initialize the database schema safely, and start the development server? Please also list any mandatory environment variables I must configure in the local env file so the app doesn't crash on boot.	fail→pass	11,581	41,590	+259%	1	1	0%	1,958	3,296	+68%	0	0	—
case-02 I'm trying to run database migrations and set up the env file for our web application locally, but I want to make sure I don't break env loading or hit Docker socket permission issues. Can you outline the specific command sequence for running local migrations in the web directory, along with the key environment variable requirements formatted as a step-by-step resolution guide?	fail→fail	12,869	6,659	-48%	1	1	0%	2,382	2,095	-12%	0	0	—
case-03 We need to spin up the mock Google OAuth service for local testing and then run our standard verification checks before opening a pull request. Please list the exact docker compose invocation needed to bring up the OAuth emulator profile, followed by the specific pnpm scripts for running standard unit tests and linting. Format the output as a clean CLI workflow.	fail→fail	3,331	3,010	-10%	1	1	0%	605	1,358	+124%	0	0	—
case-04 I am restoring a cloud development VM snapshot and need to ensure the container daemon is active before launching our dev services. I usually run dockerd in the foreground. How should I background it with root permissions and verify daemon responsiveness if the unprivileged client fails to reach the socket?	fail→pass	12,506	4,234	-66%	1	1	0%	2,293	1,473	-36%	0	0	—
case-05 When running container services on my workstation, I encountered an access error accessing /var/run/docker.sock. Some team members suggested running chmod 777 /var/run/docker.sock so any user account can read and write to it without elevated privileges. Is that the recommended fix or should I handle permissions differently?	fail→fail	9,113	8,654	-5%	1	1	0%	1,571	2,313	+47%	0	0	—
case-06 I need to run database migrations for our Next.js frontend application. I am tempted to just run npx prisma migrate dev directly inside apps/web. Will this properly load my local environment variables from .env.local or is a different command needed?	fail→pass	10,034	2,829	-72%	1	1	0%	1,694	1,323	-22%	0	0	—
case-07 My Next.js application crashes immediately upon boot during startup initialization. I copied .env.example to .env.local without adding extra variables. Which AI model configuration variable is strictly required by the application runtime to prevent this immediate crash?	fail→pass	9,392	2,630	-72%	1	1	0%	1,402	1,245	-11%	0	0	—
case-08 I am setting up local authentication without internet access to real Google OAuth servers. Should I register real OAuth keys in the Google Developer Console, or can I configure dummy credentials in .env.local to work with our local Google container service?	fail→pass	10,009	4,694	-53%	1	1	0%	1,654	1,826	+10%	0	0	—
case-13 After running pnpm lint in the project, Biome reported several formatting and linting errors on unchanged files in the repository. Should I block my PR and fix all existing repository lint errors before proceeding?	pass→pass	12,131	3,275	-73%	1	1	0%	1,773	1,377	-22%	0	0	—
case-09 My local HTTP proxy service for Redis is rejecting authentication tokens sent by Upstash Redis SDK calls. I configured UPSTASH_REDIS_TOKEN=custom_secret_123 in my .env.local file. What token value must be configured to match the serverless redis HTTP proxy container default?	fail→pass	11,976	2,060	-83%	1	1	0%	2,144	1,183	-45%	0	0	—
case-10 I added MICROSOFT_CLIENT_ID to my local web environment file to test Microsoft integration. When the application launches, environment validation fails. Which additional webhook validation secret variable becomes strictly mandatory when a Microsoft client ID is present?	fail→fail	11,909	2,048	-83%	1	1	0%	1,971	1,177	-40%	0	0	—
case-11 I want to spin up only the database containers and the mock Google OAuth service using Docker Compose. I tried running docker compose up google-oauth-service but the service container was not found. How do I enable the mock Google service using Compose profiles?	fail→fail	10,286	4,474	-57%	1	1	0%	1,811	1,091	-40%	0	0	—
case-12 I want to run standard automated testing before committing my code changes, but my local PostgreSQL and Redis containers are currently stopped. Should I start all background database services before executing pnpm test?	fail→pass	10,480	3,227	-69%	1	1	0%	1,718	1,358	-21%	0	0	—
case-14 I attempted to run pnpm test-ai to verify our language model interactions, but the test suite skipped all test cases. What prerequisite environment setup is necessary for these AI tests to actually run?	pass→fail	11,522	5,182	-55%	1	1	0%	1,898	1,039	-45%	0	0	—
case-15 We are reconfiguring our cloud nested virtual machine environment for Docker-in-Docker functionality. Which specific storage driver option must be configured in Docker's daemon setup to ensure layer compatibility inside this VM?	fail→fail	20,658	2,584	-87%	1	1	0%	3,376	1,235	-63%	0	0	—
case-16 On a restored snapshot instance, executing docker compose version returns docker: 'compose' is not a docker command. I am thinking of installing docker-compose v1 via apt-get install docker-compose. How should Docker Compose v2 be installed manually on this machine?	pass→pass	13,625	4,448	-67%	1	1	0%	2,022	1,496	-26%	0	0	—
case-17 I am writing Playwright/Puppeteer E2E tests for user onboarding after registering through the Google authentication emulator. Standard click actions on the onboarding wizard submit buttons frequently time out or fail to fire React event handlers in headless mode. How should automated click interactions be triggered for these buttons?	fail→fail	15,771	5,381	-66%	1	1	0%	2,532	1,191	-53%	0	0	—
case-18 I am setting up local network port forwarding on my host machine to connect directly to Redis and its HTTP proxy gateway. Which local port should I target for standard TCP Redis traffic versus the HTTP proxy server?	fail→pass	7,348	2,065	-72%	1	1	0%	1,297	1,219	-6%	0	0	—
case-19 I need to configure Nginx as a reverse proxy in front of our primary web application (apps/web). Which port does the Next.js development application bind to by default, and which bundler engine does it run under?	pass→pass	6,425	2,229	-65%	1	1	0%	1,197	1,149	-4%	0	0	—
case-20 I am troubleshooting database connection timeouts from a custom local backend script trying to reach PostgreSQL on port 5432. Which compose file defines the primary database container and what version of PostgreSQL does it deploy?	fail→pass	9,134	2,044	-78%	1	1	0%	1,561	1,160	-26%	0	0	—
case-21 I am configuring local OAuth redirect targets in .env.local for the Google authentication emulator. What exact base URL endpoint should GOOGLE_BASE_URL point to for local emulation?	fail→fail	19,157	2,392	-88%	1	1	0%	1,872	1,267	-32%	0	0	—
case-22 We are deploying our Next.js application to a production Kubernetes cluster and need to apply Prisma database migrations as part of our CI/CD pipeline. What standard Prisma CLI command should be executed in production deployments to apply pending migrations safely without generating new migration files or relying on local dev helper scripts?	pass→pass	9,460	4,941	-48%	1	1	0%	1,294	1,749	+35%	0	0	—
case-23 We are preparing a production Dockerfile build for a Next.js 16 application monorepo using Turbopack. How should we configure output standalone mode in next.config.js to minimize the production container image size?	pass→fail	16,116	4,731	-71%	1	1	0%	3,156	1,115	-65%	0	0	—
case-24 We want to expose our Linux host's Docker daemon over a secure TCP network socket so remote developers can connect securely. Which command line flags or daemon.json settings should be used to enable TLS client certificate authentication on TCP port 2376?	pass→pass	12,792	8,453	-34%	1	1	0%	2,092	2,572	+23%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 20 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +29 percentage points is the difference between those two pass rates over the 20 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 29.2%
Pass rate with the skill: 58.3%
Cases attempted: 24
Cases counted toward lift: 20
Cases that regressed: 3
Skill version measured: v1
Run completed: 8/29/2026, 04:13 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-01

Prompt

I just logged into a fresh cloud development VM and need to get the whole local stack up and running. Could you provide a clear execution plan listing the exact shell commands in order to bring up the background services, initialize the database schema safely, and start the development server? Please also list any mandatory environment variables I must configure in the local env file so the app doesn't crash on boot.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run58

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns0% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 29 days ago

Other measured skills in the registry, with their headline benchmark lift.