@velaro/cli@1.4.39
Velaro Workspace v20 — command-line interface for managing bots, knowledge base ingestion, MCP API keys, search indexes, and ops health.
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThe package contains an explicit AI-agent configuration mutation that can install a capability-bearing MCP server, which requires a warning under policy. No automatic foreign AI-agent control-surface mutation, self-dependent lifecycle chain, credential exfiltration, or concealed payload execution was found.
Static reason
One or more suspicious static signals were detected.
Trigger
The user runs velaro mcp-key install.
Impact
A later Claude Code session can launch the Velaro MCP server with that key, granting the configured server access to the user's Velaro account capabilities.
Mechanism
Creates an MCP key and writes an npx-backed server entry to ~/.claude.json.
AI rationale
The package contains an explicit AI-agent configuration mutation that can install a capability-bearing MCP server, which requires a warning under policy. No automatic foreign AI-agent control-surface mutation, self-dependent lifecycle chain, credential exfiltration, or concealed payload execution was found.
Evidence
lib/commands/mcp-key.js~/.claude.json~/.claude.json.tmp-${process.pid}
Network endpoints1
api-admin-us-east.velaro.com
Decision evidence
public snapshotAI called this Suspicious at 91.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- The explicit mcp-key install command creates an API key and adds a Velaro MCP server to the user's Claude configuration.
- That MCP entry runs npx to obtain and start @velaro/mcp-server with the newly created key.
- The install lifecycle runs code that writes package-owned Velaro configuration state.
- The user-invoked update command globally installs the latest release of this package.
- The screenshot feature can use supplied environment credentials or Azure Key Vault credentials to log into Velaro staging.
Evidence against
- The Claude configuration change is confined to the explicit mcp-key install command, not the install lifecycle.
- The manifest has no runtime dependency on @velaro/cli, so there is no nested self-install lifecycle chain.
- The lifecycle hook imports only the banner routine; it does not alter AI-agent configuration or contact a network endpoint.
- The package's API calls use authenticated Velaro service endpoints for its stated CLI operations.
Behavioral surface
ChildProcessEnvironmentVars
Source & flagged code
8 flaggedpackage.jsonView file
•scripts.postinstall = node scripts/postinstall.js
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node scripts/postinstall.js
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkglib/kb-screenshot.jsView file
19import { fileURLToPath } from 'url';
L20: import { execFile } from 'child_process';
L21: import { promisify } from 'util';
High
Child Process
Package source references child process execution.
lib/kb-screenshot.jsView on unpkg · L19Findings
8 High4 Medium5 Low
HighInstall Time Lifecycle Scriptspackage.json
HighChild Processlib/kb-screenshot.js
HighShell
HighSame File Env Network Executionlib/kb-screenshot.js
HighCommand Output Exfiltrationlib/commands/update.js
HighTrigger Reachable Command Output Exfiltrationlib/commands/update.js
HighRuntime Package Installlib/commands/update.js
HighSemantic Analysis Limitedlib/kb-screenshot.js#virtual:normalized:round1
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
Affected versions and remediation
This report applies to @velaro/cli@1.4.39.
See all versions of @velaro/cli for other recorded verdicts.
- Review the evidence and your use of @velaro/cli@1.4.39 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.