Summary
@tencent-ai/agent-sdk v0.3.91 was classified as CRITICAL RISK with a risk score of 12835.5. Sigil detected 861 findings across 122 files, covering phases including network exfiltration, install hooks, code patterns, obfuscation, provenance. Review the findings below before installing this package.
v0.3.91
22 March 2026, 19:32 UTC
by Sigil Bot
Risk Score
12835.5
Findings
861
Files Scanned
122
Provenance
Findings by Phase
Phase Ordering
Phases are ordered by criticality, with the most dangerous at the top. Click any phase header to expand or collapse its findings. Critical phases are expanded by default.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/codebuddy-headless.js:1
!function(D,ei){if("object"==typeof exports&&"object"==typeof module)module.exports=ei();else if("function"==typeof define&&define.amd)define([],ei);else{var ea=ei();for(var el in ea)("object"==typeof exports?exports:D)[el]=ea[el]}}(global,()=>(()=>{var __webpack_modules__={80707(D,ei,ea){"use strict";var el=this&&this.__createBinding||(Object.create?function(D,ei,ea,el){void 0===el&&(el=ea);var ec=Object.getOwnPropertyDescriptor(ei,ea);(!ec||("get"in ec?!ei.__esModule:ec.writable||ec.configurabWhy was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/codebuddy-headless.js:1
!function(D,ei){if("object"==typeof exports&&"object"==typeof module)module.exports=ei();else if("function"==typeof define&&define.amd)define([],ei);else{var ea=ei();for(var el in ea)("object"==typeof exports?exports:D)[el]=ea[el]}}(global,()=>(()=>{var __webpack_modules__={80707(D,ei,ea){"use strict";var el=this&&this.__createBinding||(Object.create?function(D,ei,ea,el){void 0===el&&(el=ea);var ec=Object.getOwnPropertyDescriptor(ei,ea);(!ec||("get"in ec?!ei.__esModule:ec.writable||ec.configurabWhy was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/codebuddy-headless.js:32
`,el=D.write(ea);if(D.flush?.(),!el){if(D.destroyed||D.writableEnded)return;try{await (0,tg.once)(D,"drain")}catch{}}}clearInFlightPrompt(D,ei){D.inFlightPrompts.delete(ei)&&this.clearPromptTimeout(ei)}scheduleStreamClose(D,ei,ea,el){this.logger.info?.(`[ACP StreamManager] scheduleStreamClose: Scheduling delayed close for stream ${ea} (${ek.STREAM_CLOSE_DELAY_MS}ms)`);let ec=setTimeout(()=>{this.streamCloseTimers.delete(ea),this.logger.info?.(`[ACP StreamManager] scheduleStreamClose: Delayed clWhy was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/codebuddy-headless.js:604
`,{mode:eL.ConsoleOutputMode.NATIVE})});try{await ei.start(),await new Promise(D=>{process.on("SIGINT",async()=>{this.consoleManager.log("\nShutting down...\n",{mode:eL.ConsoleOutputMode.NATIVE}),await ei.stop(),D(),process.exit(0)}),process.on("SIGTERM",async()=>{await ei.stop(),D(),process.exit(0)})})}catch(D){this.consoleManager.log(`Failed to start remote control: ${D.message}
`,{mode:eL.ConsoleOutputMode.NATIVE}),process.exit(1)}}};remote_control_mode_strategy_decorate([(0,e$.Autowired)(eL.Why was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/codebuddy-headless.js:738
esac
`}async setup(D,ei){let ea=this.rcsToUpdate(D),el=this.envScript(D,ei);for(let ei of ea)await appendIfNotExists(D,ei,el)}};eG=shell_setup_decorate([(0,ey.Component)(eC)],eG);let eZ=class{get name(){return"bash"}async exists(D){return(await this.rcsToUpdate(D)).length>0}rcfiles(D){return[".bash_profile",".bash_login",".bashrc"].map(ei=>(0,eP.join)(D.homedir(),ei))}async rcsToUpdate(D){let{access:ei}=D.fs.promises;return filterAsync(this.rcfiles(D),D=>ei(D).then(()=>!0).caWhy was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/codebuddy-headless.js:1294
You may need to restart Warp for changes to take effect`)+`
See ${D}`}catch(D){throw Error(`Failed to install Warp Shift+Enter key binding: ${D}`)}}static async setup(D){switch(D){case eU.TerminalType.ITERM:case eU.TerminalType.ITERM_LEGACY:return await this.setupITerm2();case eU.TerminalType.APPLE_TERMINAL:return await this.setupAppleTerminal();case eU.TerminalType.VSCODE:return await this.setupVSCode("VSCode");case eU.TerminalType.CODEBUDDY:return await this.setupVSCode("CodeBuddy");case eU.TeWhy was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
install-pip-setup-exec
CRITICALsetup.py executes code at install time
package/cli/dist/web-ui/assets/index-B-fY-Ahr.js:381
WARNING: This link could potentially be dangerous`)){const p=window.open();if(p){try{p.opener=null}catch{}p.location.href=c}else console.warn("Opening link blocked as opener could not be cleared")}}n.OscLinkProvider=v=d([g(0,_.IBufferService),g(1,_.IOptionsService),g(2,_.IOscLinkService)],v)},6193:(b,n)=>{Object.defineProperty(n,"__esModule",{value:!0}),n.RenderDebouncer=void 0,n.RenderDebouncer=class{constructor(o,d){this._renderCallback=o,this._coreBrowserService=d,this._refreshCallbacks=[]}dWhy was this flagged?
This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.
Badge
Markdown
[](https://sigilsec.ai/scans/1D8F7677-E1E8-4D59-9DE7-46D0F14C679D)HTML
<a href="https://sigilsec.ai/scans/1D8F7677-E1E8-4D59-9DE7-46D0F14C679D"><img src="https://sigilsec.ai/badge/npm/@tencent-ai/agent-sdk" alt="Sigil Scan"></a>Run This Scan Yourself
Scan your own packages
Run Sigil locally to audit any package before it touches your codebase.
Early Access
Get cloud scanning, threat intel, and CI/CD integration.
Join 150+ developers on the waitlist.
Get threat intelligence and product updates
Security research, new threat signatures, and product updates. No spam.
Other npm scans
Believe this result is incorrect? Request a review or see our Terms of Service and Methodology.