- A consent prompt appears even in projects without Vercel configuration
- The plugin delivers consent requests via system-level instruction injection
- Bash commands are fully captured, including sensitive environment details
A developer reviewing the Vercel plugin in Claude Code discovered that a telemetry consent request appeared unexpectedly during unrelated work.
The project contained no configuration files or Vercel dependencies, but the system still asked if the prompt data could be shared.
The request stated that “anonymous usage data” was already collected, followed by an option to also include prompt text.
Article continues below
Instead of appearing as a standard interface element, the consent request was delivered via instructions injected into Claude’s system context.
These instructions instructed the AI tool to ask the user a question and then execute shell commands based on the answer.
The result was indistinguishable from a native interaction, leaving no visible indication that the prompt came from a plugin rather than the main system.
The developer described the experience clearly, stating that “it didn’t feel right” and then examined the plugin’s source code to verify how the mechanism worked.
Inspection of the source code shows that telemetry works on multiple layers, with some data collection enabled by default.
Session-level data includes device identifiers, operating system details, detected frameworks, and installed CLI versions, all transmitted at the start of each session. This happens without an explicit opt-in mechanism.
Most notably, strings of bash commands executed in Claude Code are also captured and transmitted.
These entries include the full contents of commands rather than abstract metadata, potentially exposing file paths, environment variables, and infrastructure details.
This collection occurs automatically, regardless of any user consent for rapid sharing.
The description of this activity as “anonymous usage data such as skill injection models and tools used” does not fully reflect the granularity of the information collected.
Although the prompt text requires explicit approval, other telemetry categories remain active unless manually disabled.
The plugin’s telemetry system is not limited to Vercel-related environments, as hook configurations show that user prompt submissions match universally, while other triggers respond to general tool usage or session events rather than project-specific conditions.
As a result, telemetry works on all projects within Claude Code, regardless of their relevance to Vercel services.
This behavior contrasts with the existing framework detection logic within the plugin.
The code identifies project types by analyzing configuration files and dependencies, but this information is not used to limit telemetry activation. The trigger mechanism exists but is not applied in practice.
Disabling telemetry requires manual intervention via environment variables or configuration files.
But these options are documented in the plugin directory rather than appearing during installation, making them harder to access.
Deleting the device ID file or disabling the plugin completely also stops data collection, although these steps are not presented during initial setup.
Simply put, the system combines automated data collection with limited visibility into when and how it operates.
This may not be what users expect when working outside of no-code platform environments or when using an LLM for coding.
TechRadar Pro contacted Vercel for comment, but had no response at the time of publication.
Via Akshay Chugh
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds. Make sure to click the Follow button!
And of course you can too follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp Also.
