registry / wechat-bro / 1.6.4
wechat-bro@1.6.4
AI agent interface for WeChat Web — send/receive messages, manage contacts, upload media, transcribe voice
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThis is a dangerous, user-invoked AI automation capability rather than a confirmed malicious package. The postinstall Chrome download is package-aligned but adds avoidable install-time network and filesystem effects.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
The user runs wechat-bro orchestrator and a watched message or WebSocket command arrives.
Impact
A malicious message can influence a broadly privileged local AI session and cause unintended local actions or WeChat replies.
Mechanism
AI-agent task dispatch with permission-bypassing harnesses and workspace instruction-file writes.
AI rationale
This is a dangerous, user-invoked AI automation capability rather than a confirmed malicious package. The postinstall Chrome download is package-aligned but adds avoidable install-time network and filesystem effects.
Evidence
package.jsonsrc/orchestrator.jssrc/ws-server.js~/.wechat-bro/chromiumAGENTS.md.task-*.md
Network endpoints2
wx.qq.comfile.wx.qq.com
Decision evidence
public snapshotAI called this Suspicious at 90.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The install hook runs npx to download Chrome into the user's home directory.
- The explicit orchestrator mode replaces or creates AGENTS.md in each agent workspace.
- Built-in agent harnesses launch Codex with danger-full-access and other CLIs with permission-bypass options.
- Its WebSocket server dispatches received commands without enforcing the optional identity command.
Evidence against
- The risky orchestration path requires an explicit wechat-bro orchestrator command.
- No self-dependency, hidden payload, credential harvesting, or attacker-controlled exfiltration endpoint was found.
- WeChat uploads use the logged-in session parameters for the stated media-upload feature.
Behavioral surface
ChildProcessEnvironmentVarsFilesystemNetworkWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
7 flaggedpackage.jsonView file
•scripts.postinstall = node -e "const p=require('path'),o=require('os'),d=p.join(o.homedir(),'.wechat-bro','chromium');process.env.PUPPETEER_CACHE_DIR=d;const pup=require('puppeteer');(async()=>{try{awai...
Critical
Red Install Lifecycle Script
Install-time lifecycle script matches a deterministic static-gate block pattern.
package.jsonView on unpkg•scripts.postinstall = node -e "const p=require('path'),o=require('os'),d=p.join(o.homedir(),'.wechat-bro','chromium');process.env.PUPPETEER_CACHE_DIR=d;const pup=require('puppeteer');(async()=>{try{awai...
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1
High
Semantic Analysis Limited
A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
package.jsonView on unpkgFindings
2 Critical5 High3 Medium4 Low
CriticalRed Install Lifecycle Scriptpackage.json
CriticalPrevious Version Dangerous Deltasrc/cli.js
HighInstall Time Lifecycle Scriptspackage.json
HighChild Processsrc/upload.js
HighSame File Env Network Executionsrc/cli.js
HighRuntime Package Installsrc/cli.js
HighSemantic Analysis Limitedpackage.json
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to wechat-bro@1.6.4.
See all versions of wechat-bro for other recorded verdicts.
- Review the evidence and your use of wechat-bro@1.6.4 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.