@fragment-dev/cli@2026.10.28-5
FRAGMENT CLI
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe CLI sends runtime usage analytics by default, and the inspected package source does not clearly disclose that collection. The event contains limited command and runtime metadata, so this is suspicious telemetry rather than evidence of credential theft or malicious exfiltration.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
Running the `fragment` CLI; collection can be disabled with `DISABLE_CLI_ANALYTICS=true`.
Impact
Command usage and limited runtime metadata are reported to the analytics service; no credential or project-data exfiltration was identified.
Mechanism
The CLI captures a PostHog event containing the command ID and basic runtime properties.
AI rationale
The CLI sends runtime usage analytics by default, and the inspected package source does not clearly disclose that collection. The event contains limited command and runtime metadata, so this is suspicious telemetry rather than evidence of credential theft or malicious exfiltration.
Evidence
dist/chunk-VPOARVKJ.js
Network endpoints1
us.i.posthog.com
Decision evidence
public snapshotAI called this Suspicious at 83.0% confidence as Unknown with medium false-positive risk.
Evidence for warning
- The `fragment` executable starts the oclif CLI, whose command initialization sends an analytics event by default unless `DISABLE_CLI_ANALYTICS=true`.
- The event reports the command ID and basic runtime details to PostHog; the inspected payload contains no credentials or user/project data.
Evidence against
- The manifest declares no install lifecycle scripts.
- The analytics code includes an environment variable that disables collection.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetwork
HighEntropyStringsPossibleObfuscationTelemetryUrlStrings
Source & flagged code
7 flaggeddist/open-SJMVVA3H.jsView file
11import { promisify as promisify5 } from "node:util";
L12: import childProcess from "node:child_process";
L13: import fs5, { constants as fsConstants2 } from "node:fs/promises";
High
Child Process
Package source references child process execution.
dist/open-SJMVVA3H.jsView on unpkg · L11dist/chunk-VPOARVKJ.jsView file
105Cross-file remote execution chain: dist/chunk-VPOARVKJ.js spawns dist/chunk-QF5EN33L.js; helper contains network access plus dynamic code execution.
L105: function isProd() {
L106: return !["development", "test"].includes(process.env.NODE_ENV ?? "");
L107: }
...
L2227: if (useCache && cache.has(path)) {
L2228: return JSON.parse(cache.get(path));
L2229: }
...
L2308:
L2309: // ../../node_modules/@oclif/core/package.json
L2310: var require_package = __commonJS({
...
L2316: author: "Salesforce",
L2317: bugs: "https://github.com/oclif/core/issues",
L2318: dependencies: {
High
Cross File Remote Execution Context
Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.
dist/chunk-VPOARVKJ.jsView on unpkg · L105Findings
1 Critical3 High5 Medium6 Low
CriticalPrevious Version Dangerous Deltadist/chunk-YKGESMWT.js
HighChild Processdist/open-SJMVVA3H.js
HighCross File Remote Execution Contextdist/chunk-VPOARVKJ.js
HighSemantic Analysis Limiteddist/chunk-QDAKBUNO.js#virtual:normalized:round1
MediumDynamic Requiredist/chunk-VPOARVKJ.js
MediumNetwork
MediumEnvironment Vars
MediumShips Build Helperbin/dev.cmd
MediumStructural Risk Force Deep Review
LowScripts Present
LowEvaldist/chunk-VPOARVKJ.js
LowFilesystem
LowHigh Entropy Strings
LowTelemetry
Affected versions and remediation
This report applies to @fragment-dev/cli@2026.10.28-5.
See all versions of @fragment-dev/cli for other recorded verdicts.
- Review the evidence and your use of @fragment-dev/cli@2026.10.28-5 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.