How do I install the Rosetta Test Setup skill?

Run npx skillmds@latest add arize-ai/rosetta-test-setup in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Rosetta Test Setup safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Rosetta Test Setup?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Rosetta Test Setup free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Rosetta Test Setup?

arize-ai (@arize-ai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Rosetta Test Setup by arize-ai — Agent Skill

Name: Rosetta Test Setup
Author: arize-ai

Rosetta Test Setup

Provision a fresh Rosetta Stone test project on Arize AX or Phoenix. Validates the target framework directory and credentials, mints a unique project name, writes a sibling .env.test-local overlay so the real .env.local is untouched, and pre-creates the AX project (Phoenix auto-creates on first trace). Part of the rosetta-test e2e flow; can also be invoked standalone.

arize-ai 2e59b59 3.6 KB Updated Sep 17, 2026

File contents

Rosetta Test — Setup Phase

Inputs

<framework> — directory name under the chosen platform
<platform> — ax or phoenix

Steps

1. Validate the target directory

PROJECT_DIR="$(pwd)/<platform>/<framework>"
test -d "$PROJECT_DIR" || { echo "No such directory: $PROJECT_DIR"; exit 1; }
test -f "$PROJECT_DIR/package.json" || { echo "Missing package.json"; exit 1; }

Check the required npm scripts. Use jq if available, otherwise node -e:

HAS_SYN=$(node -e "console.log(!!require('$PROJECT_DIR/package.json').scripts?.['synthetic-requests'])")
[ "$HAS_SYN" = "true" ] || { echo "package.json missing 'synthetic-requests' script"; exit 1; }

For phoenix platform only, also require evals:

if [ "<platform>" = "phoenix" ]; then
  HAS_EVALS=$(node -e "console.log(!!require('$PROJECT_DIR/package.json').scripts?.evals)")
  [ "$HAS_EVALS" = "true" ] || { echo "phoenix tier missing 'evals' script"; exit 1; }
fi

If validation fails, list what's actually available under that platform so the user can spot a typo:

ls "$(pwd)/<platform>/"

2. Validate credentials

Source .env.local from the target directory and confirm the platform-specific keys are non-empty:

set -a
[ -f "$PROJECT_DIR/.env.local" ] && source "$PROJECT_DIR/.env.local"
set +a

AX: require ARIZE_SPACE_ID and ARIZE_API_KEY
Phoenix: require PHOENIX_API_KEY and PHOENIX_COLLECTOR_ENDPOINT

Abort with a clear message if missing. Do not print the values themselves.

3. Generate a unique project name

TIMESTAMP=$(date -u +%Y%m%d%H%M)
RAND=$(openssl rand -hex 2)
PROJECT_NAME="rosetta-e2e-<framework>-${TIMESTAMP}-${RAND}"

Phoenix rejects project names containing /, ?, or # — the above is safe.

4. Pre-create the AX project (AX only)

ax projects create --name "$PROJECT_NAME" --space "$ARIZE_SPACE_ID"

Skip this step for Phoenix — projects auto-create on first trace ingestion.

If creation fails (rare, e.g. transient network), abort. Do not retry — let the orchestrator decide.

5. Write the env overlay

The overlay loads on top of .env.local so the real file is never mutated. Each tier's synthetic-requests and evals scripts already do --env-file-if-exists on .env and .env.local only, so the traces and evals skills must source .env.test-local explicitly into their parent shell.

ENV_OVERLAY="$PROJECT_DIR/.env.test-local"

cat > "$ENV_OVERLAY" <<EOF
# Rosetta E2E test overlay — safe to delete
# Generated $(date -u +%Y-%m-%dT%H:%M:%SZ)
EOF

if [ "<platform>" = "ax" ]; then
  echo "ARIZE_PROJECT_NAME=$PROJECT_NAME" >> "$ENV_OVERLAY"
else
  echo "PHOENIX_PROJECT_NAME=$PROJECT_NAME" >> "$ENV_OVERLAY"
fi

6. Emit state for downstream phases

Print these on their own lines, prefixed clearly, so the orchestrator (or a human resuming manually) can capture them:

PROJECT_NAME=<value>
PROJECT_DIR=<value>
PLATFORM=<value>
ENV_OVERLAY=<value>

Idempotency

If .env.test-local already exists, do not overwrite. Abort with a message: "Existing test overlay found at . Run cleanup first, or pass --keep was used previously." This protects against accidental clobbering of an in-flight run.

---name: rosetta-test-setupdescription: Provision a fresh Rosetta Stone test project on Arize AX or Phoenix. Validates the target framework directory and credentials, mints a unique project name, writes a sibling .env.test-local overlay so the real .env.local is untouched, and pre-creates the AX project (Phoenix auto-creates on first trace). Part of the rosetta-test e2e flow; can also be invoked standalone.---# Rosetta Test — Setup Phase## Inputs- `<framework>` — directory name under the chosen platform- `<platform>` — `ax` or `phoenix`## Steps### 1. Validate the target directory```bashPROJECT_DIR="$(pwd)/<platform>/<framework>"test -d "$PROJECT_DIR" || { echo "No such directory: $PROJECT_DIR"; exit 1; }test -f "$PROJECT_DIR/package.json" || { echo "Missing package.json"; exit 1; }```Check the required npm scripts. Use `jq` if available, otherwise `node -e`:```bashHAS_SYN=$(node -e "console.log(!!require('$PROJECT_DIR/package.json').scripts?.['synthetic-requests'])")[ "$HAS_SYN" = "true" ] || { echo "package.json missing 'synthetic-requests' script"; exit 1; }```For `phoenix` platform only, also require `evals`:```bashif [ "<platform>" = "phoenix" ]; then HAS_EVALS=$(node -e "console.log(!!require('$PROJECT_DIR/package.json').scripts?.evals)") [ "$HAS_EVALS" = "true" ] || { echo "phoenix tier missing 'evals' script"; exit 1; }fi```If validation fails, list what's actually available under that platform so the user can spot a typo:```bashls "$(pwd)/<platform>/"```### 2. Validate credentialsSource `.env.local` from the target directory and confirm the platform-specific keys are non-empty:```bashset -a[ -f "$PROJECT_DIR/.env.local" ] && source "$PROJECT_DIR/.env.local"set +a```- AX: require `ARIZE_SPACE_ID` and `ARIZE_API_KEY`- Phoenix: require `PHOENIX_API_KEY` and `PHOENIX_COLLECTOR_ENDPOINT`Abort with a clear message if missing. Do **not** print the values themselves.### 3. Generate a unique project name```bashTIMESTAMP=$(date -u +%Y%m%d%H%M)RAND=$(openssl rand -hex 2)PROJECT_NAME="rosetta-e2e-<framework>-${TIMESTAMP}-${RAND}"```Phoenix rejects project names containing `/`, `?`, or `#` — the above is safe.### 4. Pre-create the AX project (AX only)```bashax projects create --name "$PROJECT_NAME" --space "$ARIZE_SPACE_ID"```Skip this step for Phoenix — projects auto-create on first trace ingestion.If creation fails (rare, e.g. transient network), abort. Do not retry — let the orchestrator decide.### 5. Write the env overlayThe overlay loads on top of `.env.local` so the real file is never mutated. Each tier's `synthetic-requests` and `evals` scripts already do `--env-file-if-exists` on `.env` and `.env.local` only, so the traces and evals skills must source `.env.test-local` explicitly into their parent shell.```bashENV_OVERLAY="$PROJECT_DIR/.env.test-local"cat > "$ENV_OVERLAY" <<EOF# Rosetta E2E test overlay — safe to delete# Generated $(date -u +%Y-%m-%dT%H:%M:%SZ)EOFif [ "<platform>" = "ax" ]; then echo "ARIZE_PROJECT_NAME=$PROJECT_NAME" >> "$ENV_OVERLAY"else echo "PHOENIX_PROJECT_NAME=$PROJECT_NAME" >> "$ENV_OVERLAY"fi```### 6. Emit state for downstream phasesPrint these on their own lines, prefixed clearly, so the orchestrator (or a human resuming manually) can capture them:```PROJECT_NAME=<value>PROJECT_DIR=<value>PLATFORM=<value>ENV_OVERLAY=<value>```## IdempotencyIf `.env.test-local` already exists, do **not** overwrite. Abort with a message: "Existing test overlay found at <path>. Run cleanup first, or pass --keep was used previously." This protects against accidental clobbering of an in-flight run.

Frequently asked questions