Warn by default; block when configured. The install changes an AI agent’s MCP configuration without explicit user consent, enabling the package’s server when Claude Code restarts.
AI security review
scanned · by lpm-firewall-ai
The default postinstall hook writes an MCP server entry into the consuming project’s Claude Code configuration. This is an unconsented install-time change to an AI-agent control surface, so the package is malicious under the stated policy.
Package installation, unless the user sets the opt-out variable.
Impact
The install changes an AI agent’s MCP configuration without explicit user consent, enabling the package’s server when Claude Code restarts.
Mechanism
The hook writes an exosphere server entry to the project’s .mcp.json.
Policy narrative
During installation, the package edits the consuming project’s .mcp.json to register its MCP server, thereby changing Claude Code’s configuration without an explicit request. The hook supports an opt-out, but the configuration change is enabled by default.
Evidence
package.jsonscripts/postinstall-mcp.js.mcp.json
Decision evidence
public snapshot
AI identified this as a dangerous AI-agent capability at 99.0% confidence with low false-positive risk.
Evidence for policy risk
The package runs scripts/postinstall-mcp.js during installation.
The hook locates the consuming project’s root and targets its .mcp.json, adding a command that launches this package’s MCP server.
The hook writes that configuration automatically and tells the user to restart Claude Code to activate it.
Evidence against
The hook offers an opt-out environment variable, but setup runs by default.
The MCP server uses standard input/output transport when launched.
2In order to be iterable, non-array objects must have a [Symbol.iterator]() method.`)}var d,g=!0,C=!1;return{s:function(){a=a.call(i)},n:function(){var A=a.next();return g=A.done,A}...
Avoid installing @boomi/exosphere@8.2.0. Remove it from direct dependencies and check your lockfile for transitive copies.
Choose an independently verified alternative or release. This report does not establish that other versions are safe.
If this version ran, check your AI agents' configuration and instruction files (the files this report lists) for changes you did not make, and remove them. This finding does not by itself establish malware intent.