Security audit
Security checks for vulnerabilities and agentic risk
This is a mostly coherent Ned analytics connector, but it should be reviewed because broad prompts can make an agent use your API key to pull sensitive business and customer analytics.
Install only if you intend your agent to query Ned for Shopify analytics and you trust Ned with that data. Use a scoped or revocable API key if available, avoid sharing the key in logs or chats, and consider requiring explicit confirmation before customer-level, product-margin, churn-risk, or ad-performance queries.
The skill invokes shell-based network access via curl and a bundled bash script, but it does not declare any tool restrictions or allowed tools in metadata. That increases the chance an agent can execute shell commands unexpectedly when the skill is auto-selected, expanding the attack surface and weakening least-privilege controls.
The description contains very broad trigger phrases like profit, revenue, sales, customers, ad performance, and 'how is my store doing,' which overlap with many normal ecommerce conversations. In agent environments with automatic skill routing, this can cause over-invocation and unintended access to external services and sensitive analytics data.
The skill states that it gives the agent direct access to profitability, customer segments, churn risk, and ad efficiency data stored in Ned, but it does not prominently warn that this information will be sent to and processed by a third-party service. Because the data includes sensitive business and potentially customer-related analytics, users may unknowingly authorize external disclosure.
This skill is explicitly designed to send authenticated requests to api.meetned.com using a bearer API key, which constitutes external transmission of sensitive store analytics. While expected for functionality, it is still a real security concern because compromise, misconfiguration, or unexpected invocation could disclose confidential business metrics and customer-related insights to a third party.
https://api.meetned.com/api/v1
Auth: Authorization: Bearer $NED_API_KEY
The example curl command demonstrates outbound authenticated transmission of profitability data to a third-party API. Although this is core functionality, embedding executable shell snippets increases the chance an agent will perform network actions automatically without sufficient guardrails or user awareness.
# Quick profit check
curl -s -H "Authorization: Bearer $NED_API_KEY" \
"https://api.meetned.com/api/v1/profitability/summary?period=today"
# Top products by profit
curl -s -H "Authorization: Bearer $NED_API_KEY" \
This example transmits product profitability data, which can reveal commercially sensitive margins, COGS, and performance information, to an external endpoint using a bearer token. The risk is amplified in an agent setting because shell-capable skills may execute these calls with minimal user visibility.
# Top products by profit
curl -s -H "Authorization: Bearer $NED_API_KEY" \
"https://api.meetned.com/api/v1/profitability/products?period=last_30_days"
# At-risk whale customers
curl -s -H "Authorization: Bearer $NED_API_KEY" \
This request accesses customer-summary data including at-risk whales, which may involve sensitive customer segmentation and behavioral analytics, and sends/receives it through a third-party API. Customer-related analytics are more privacy-sensitive than aggregate metrics, so unintended or opaque transmission poses greater confidentiality and compliance risk.
curl -s -H "Authorization: Bearer $NED_API_KEY"
"https://api.meetned.com/api/v1/customers/summary?period=last_90_days" | jq '.data.at_risk_whales'
## Query Script
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
exit 1
fi
URL="https://api.meetned.com/api/v1/${ENDPOINT}?period=${PERIOD}"
response=$(curl -s -w "\n%{http_code}" -H "Authorization: Bearer $NED_API_KEY" "$URL")
http_code=$(echo "$response" | tail -1)
No suspicious patterns detected.