registry / AI-agent capability abuse / @knockdata/objectexplorer / 0.8.0

@knockdata/objectexplorer@0.8.0

The VSCode for Cloud Storage — Explore, Analyze, Search your cloud storage locally with single click

AI Security Review

scanned 24d ago · by lpm-firewall-ai

The package offers explicit AI-agent configuration mutation through its server API. That capability warrants a warning, but the inspected source does not establish malware or an unconsented install-time attack.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review

Trigger

A client sends the server's MCP installation request with agent selection, URL, and enabled flag.

Impact

A user who invokes this capability can add a remote MCP server to their Codex or Claude Code configuration.

Mechanism

The handler writes an objectexplorer MCP configuration block using the supplied URL and a server token.

AI rationale

Evidence

.codex/config.toml.claude.json

Decision evidence

public snapshot

AI called this Suspicious at 86.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

The manifest exposes a command-line launcher and server entrypoint, with no lifecycle scripts.
The server accepts an installation request for a supported agent and passes a caller-provided URL and generated token to the agent configuration writer.
The Codex writer creates or replaces an objectexplorer MCP server block containing that URL and bearer token.
This is an explicit server API action, not an npm install hook, but it changes a user-level AI-agent configuration.

Evidence against

No package lifecycle hook was found.
No fixed external endpoint or secret-export sink was identified in the inspected installation path.

Behavioral surface

Source

ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkShell

Supply chain

HighEntropyStringsMinifiedPossibleObfuscationTelemetry

Source & flagged code

27 flagged

server/WebServer.mjsView file

181`+Ee:"",Ge!==null?` L182: `+Ge:""))}}}return z}a(UI,"constructClassInstance");function $I(T,O,H){{var Z=_n(O)||"Component",b=T.render;b||(O.prototype&&typeof O.prototype.render=="function"?o("%s(...): No `r... L183: ... L194: Error generating stack: `+Z.message+` L195: `+Z.stack}}a(NS,"getStackByComponentStackNode");var w0=n.ReactCurrentDispatcher,Vc=n.ReactDebugCurrentFrame,zg=0,bd=1,kS=2,g3=3,LS=4,lf=0,_3=1,cf=2,T0=12800;function b3(T){return c... L196: `));if(!R)return null;ge=Le(R,!0)}return ge.shift()},"FS_stdin_getChar"),nt={ttys:[],init(){},shutdown(){},register(R,P){nt.ttys[R]={input:[],output:[],ops:P},V.registerDevice(R,nt... L197: `)){let n=r.trim();if(n.startsWith(`${t} `)||n.startsWith(`${t} `))return n.slice(t.length).trim()}return""}function ont(e){let t=e.match(/^loc

Critical

Same File Env Network Execution

A single source file combines environment access, network access, and code or shell execution with blocking evidence.

server/WebServer.mjsView on unpkg · L181

1import { createRequire as __cr } from 'module'; const require = __cr(import.meta.url); L2: var sMe=Object.create;var L3=Object.defineProperty;var lMe=Object.getOwnPropertyDescriptor;var cMe=Object.getOwnPropertyNames;var uMe=Object.getPrototypeOf,dMe=Object.prototype.has... L3: `)}catch(n){process.stderr.write(` L4: logger: could not write ${t_} (${n.code})`)}}function Pr(...e){let t=P3(2,"info");process.stderr.write(` ... L9: `+MZ(t,i.offset,i.len))}return n.value}function MZ(e,t,r=1){return`${e.replace(/(\t)|(\r)|(\n)/g,(n,i,o)=>i?"\u2409":o?"\u240D":"\u240A")} L10: ${"".padEnd(t)}${"^".repeat(r)}`}function GZ(e){return zFe(HN,e)}function Yl(e){if(!e.type)throw new Error("This is not an AST node.");switch(e.type){case"universal":return Y3(e.na... L11: `.repeat(o)+t:(e.rawText=t,e.lea

Critical

Hardcoded Runtime Data Exfiltration

Source sends credentials or rich application records to a package-controlled external receiver enabled by default.

server/WebServer.mjsView on unpkg · L1

Critical

Credential Exfiltration

Source appears to send environment or credential material to an external endpoint.

server/WebServer.mjsView on unpkg · L1

Critical

Browser Credential Phishing

Source appears to collect browser login credentials for exfiltration.

server/WebServer.mjsView on unpkg · L1

1Trigger-reachable chain: manifest.main -> server/WebServer.mjs L1: import { createRequire as __cr } from 'module'; const require = __cr(import.meta.url); L2: var sMe=Object.create;var L3=Object.defineProperty;var lMe=Object.getOwnPropertyDescriptor;var cMe=Object.getOwnPropertyNames;var uMe=Object.getPrototypeOf,dMe=Object.prototype.has... L3: `)}catch(n){process.stderr.write(` L4: logger: could not write ${t_} (${n.code})`)}}function Pr(...e){let t=P3(2,"info");process.stderr.write(` ... L9: `+MZ(t,i.offset,i.len))}return n.value}function MZ(e,t,r=1){return`${e.replace(/(\t)|(\r)|(\n)/g,(n,i,o)=>i?"\u2409":o?"\u240D":"\u240A")} L10: ${"".padEnd(t)}${"^".repeat(r)}`}function GZ(e){return zFe(HN,e)}function Yl(e){if(!e.type)throw new Error("This is not an AST node.");switch(e.type){case"uni

Critical

Trigger Reachable Dangerous Capability

A package entrypoint or install-time lifecycle script reaches a source file with blocking dangerous behavior.

server/WebServer.mjsView on unpkg · L1

196`));if(!R)return null;ge=Le(R,!0)}return ge.shift()},"FS_stdin_getChar"),nt={ttys:[],init(){},shutdown(){},register(R,P){nt.ttys[R]={input:[],output:[],ops:P},V.registerDevice(R,nt... L197: `)){let n=r.trim();if(n.startsWith(`${t} `)||n.startsWith(`${t} `))return n.slice(t.length).trim()}return""}function ont(e){let t=e.match(/^local\[(\*|\d+)\]$/);return t?t[1]==="*"... L198: `).find(n=>n.trim().startsWith("version"));return r?r.split("=")[1].trim():""}catch{return""}}async function k4(e){let t=N4(e),r=[];try{r=await g5.readdir(t)}catch{r=[]}let n=[];fo...

High

Child Process

Package source references child process execution.

server/WebServer.mjsView on unpkg · L196

High

Shell

Package source references shell execution.

server/WebServer.mjsView on unpkg · L196

•Manifest-reachable source captures an API credential, sends it to a fixed unofficial gateway, and persists that redirection in agent or shell configuration. server/WebServer.mjs: `),"blockTransform")})}function bk(e,t){if(!t)return e;let r=typeof t[0]=="string"?t[0]:"[",n=typeof t[1]=="string"?t[1]:"]";return r+e+n}function Bee(e,t,r,n,i){let o=typeof t=="f... `);n.next.cells.push({colspan:t,rowspan:r,text:i})}closeTableRow(){let t=this._popStackItem();t.next.rows.push(t.cells)}closeTable({tableToString:t,leadingLineBreaks:r=2,trailingLi... `+Z.stack}}a(NS,"getStackByComponentStackNode");var w0=n.ReactCurrentDispatcher,Vc=n.ReactDebugCurrentFrame,zg=0,bd=1,kS=2,g3=3,LS=4,lf=0,_3=1,cf=2,T0=12800;function b3(T){return c... `));if(!R)return null;ge=Le(R,!0)}return ge.shift()},"FS_stdin_getCha

High

Credential Redirect Persistence

Manifest-reachable source sends a prompted API credential to a fixed unofficial gateway and persists the redirection.

server/WebServer.mjsView on unpkg

High

Command Output Exfiltration

Source combines command execution, command-output handling, and outbound requests; review data flow before blocking.

server/WebServer.mjsView on unpkg · L181

High

Cloud Metadata Access

Source reaches cloud instance metadata or link-local credential endpoints.

server/WebServer.mjsView on unpkg · L1

15`).map(o=>"> "+o).join(` L16: `),"blockTransform")})}function bk(e,t){if(!t)return e;let r=typeof t[0]=="string"?t[0]:"[",n=typeof t[1]=="string"?t[1]:"]";return r+e+n}function Bee(e,t,r,n,i){let o=typeof t=="f... L17: `)}splitLongWord(t){let r=[],n=0;for(;t.length>this.maxLineLength;){let i=t.substring(0,this.maxLineLength),o=t.substring(this.maxLineLength),s=i.lastIndexOf(this.wrapCharacters[n]... ... L194: Error generating stack: `+Z.message+` L195: `+Z.stack}}a(NS,"getStackByComponentStackNode");var w0=n.ReactCurrentDispatcher,Vc=n.ReactDebugCurrentFrame,zg=0,bd=1,kS=2,g3=3,LS=4,lf=0,_3=1,cf=2,T0=12800;function b3(T){return c... L196: `));if(!R)return null;ge=Le(R,!0)}return ge.shift()},"FS_stdin_getChar"),nt={ttys:[],init(){},shutdown(){},register(R,P){nt.ttys[R]={input:[],output

High

Remote Agent Bridge

Source exposes local file and command tools to a remote model endpoint.

server/WebServer.mjsView on unpkg · L15

1Cross-file remote execution chain: server/WebServer.mjs spawns app/assets/index.js; helper contains network access plus dynamic code execution. L1: import { createRequire as __cr } from 'module'; const require = __cr(import.meta.url); L2: var sMe=Object.create;var L3=Object.defineProperty;var lMe=Object.getOwnPropertyDescriptor;var cMe=Object.getOwnPropertyNames;var uMe=Object.getPrototypeOf,dMe=Object.prototype.has... L3: `)}catch(n){process.stderr.write(` L4: logger: could not write ${t_} (${n.code})`)}}function Pr(...e){let t=P3(2,"info");process.stderr.write(` ... L9: `+MZ(t,i.offset,i.len))}return n.value}function MZ(e,t,r=1){return`${e.replace(/(\t)|(\r)|(\n)/g,(n,i,o)=>i?"\u2409":o?"\u240D":"\u240A")} L10: ${"".padEnd(t)}${"^".repeat(r)}`}function GZ(e){return zFe(HN,e)}function Yl(

High

Cross File Remote Execution Context

Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.

server/WebServer.mjsView on unpkg · L1

1Trigger-reachable credential exfiltration chain: manifest.main -> server/WebServer.mjs L1: import { createRequire as __cr } from 'module'; const require = __cr(import.meta.url); L2: var sMe=Object.create;var L3=Object.defineProperty;var lMe=Object.getOwnPropertyDescriptor;var cMe=Object.getOwnPropertyNames;var uMe=Object.getPrototypeOf,dMe=Object.prototype.has... L3: `)}catch(n){process.stderr.write(` L4: logger: could not write ${t_} (${n.code})`)}}function Pr(...e){let t=P3(2,"info");process.stderr.write(` ... L9: `+MZ(t,i.offset,i.len))}return n.value}function MZ(e,t,r=1){return`${e.replace(/(\t)|(\r)|(\n)/g,(n,i,o)=>i?"\u2409":o?"\u240D":"\u240A")} L10: ${"".padEnd(t)}${"^".repeat(r)}`}function GZ(e){return zFe(HN,e)}function Yl(e){if(!e.type)throw new Error("This is not an AST node.")

High

Trigger Reachable Credential Exfiltration

A manifest entrypoint or package-local install chain reaches credential exfiltration behavior.

server/WebServer.mjsView on unpkg · L1

181Trigger-reachable command-output exfiltration chain: manifest.main -> server/WebServer.mjs L181: `+Ee:"",Ge!==null?` L182: `+Ge:""))}}}return z}a(UI,"constructClassInstance");function $I(T,O,H){{var Z=_n(O)||"Component",b=T.render;b||(O.prototype&&typeof O.prototype.render=="function"?o("%s(...): No `r... L183: ... L194: Error generating stack: `+Z.message+` L195: `+Z.stack}}a(NS,"getStackByComponentStackNode");var w0=n.ReactCurrentDispatcher,Vc=n.ReactDebugCurrentFrame,zg=0,bd=1,kS=2,g3=3,LS=4,lf=0,_3=1,cf=2,T0=12800;function b3(T){return c... L196: `));if(!R)return null;ge=Le(R,!0)}return ge.shift()},"FS_stdin_getChar"),nt={ttys:[],init(){},shutdown(){},register(R,P){nt.ttys[R]={input:[],output:[],ops:P},V.registerDevice(R,nt... L197: `)){let n=r.trim();if(n.startsWith(`${t} `)||n.start

High

Trigger Reachable Command Output Exfiltration

A manifest entrypoint or package-local install chain reaches command-output exfiltration behavior.

server/WebServer.mjsView on unpkg · L181

app/assets/index.jsView file

1764contains invisible/control Unicode U+200B (zero width space) ${g}`)).style(`margin-left`,u?`${+u}px`:null).style(`width`,p===void 0?null:`${+p}px`).style(`font-variant`,I$(o,`normal`)).call(z$,f).node()}var u0=new Map([[`symbol`,Ket],[`color`,f0],[`opacity`,Jet]]);function qet(e,t,n={}){return(r,a)=>

Critical

Trojan Source Unicode

Source contains bidi control or invisible Unicode characters associated with Trojan Source attacks.

app/assets/index.jsView on unpkg · L1764

1547`||r===`\r`?(yield[t,n,!0],n+=r===`\r`&&e[n+1]===` L1548: `?2:1,t=n):n+=1}yield[t,n,!0]}function hA(e,t,n){return e[n-1]===fA?e.slice(t,n)+`-`:e.slice(t,n)}function HBe(e,t=1/0){if(t===1/0)return e.split(/\r\n?|\n/g);{let n=t*100,r=[],a,o... L1549: `)}function GWe(e,t,n){let r=[`x: ${t.source}`];n&&r.push(`y: ${n.source}`),r.push(`tip: true`);let a=`{${r.join(`, `)}}`,o=[];return e.x.reduce!==null&&o.push(`x: ${XN(e.x.reduce)...

High

Eval

Package source references dynamic code evaluation.

app/assets/index.jsView on unpkg · L1547

app/assets/esm-worker.jsView file

•matchType = normalized_sha256 matchedPackage = @knockdata/objectexplorer@0.7.0 matchedPath = app/assets/esm-worker.js matchedIdentity = npm:[redacted]:0.7.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

app/assets/esm-worker.jsView on unpkg

1const e=`https://unpkg.com/@ffmpeg/core@0.12.9/dist/umd/ffmpeg-core.js`;var t;(function(e){e.LOAD=`LOAD`,e.EXEC=`EXEC`,e.FFPROBE=`FFPROBE`,e.WRITE_FILE=`WRITE_FILE`,e.READ_FILE=`RE...

Medium

Dynamic Require

Package source references dynamic require/import behavior.

app/assets/esm-worker.jsView on unpkg · L1

app/assets/lightgbmc.wasmView file

•path = app/assets/lightgbmc.wasm kind = wasm_module sizeBytes = 21264 magicHex = [redacted]

Medium

Ships Wasm Module

Package ships WebAssembly modules.

app/assets/lightgbmc.wasmView on unpkg

server/kernel.pyView file

•path = server/kernel.py kind = build_helper sizeBytes = 10781 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

server/kernel.pyView on unpkg

demo/ebook/the-lighthouse-keeper.epubView file

•path = demo/ebook/the-lighthouse-keeper.epub kind = high_entropy_blob sizeBytes = 33071 magicHex = [redacted]

High

Ships High Entropy Blob

Package ships high-entropy non-source blobs.

demo/ebook/the-lighthouse-keeper.epubView on unpkg

•path = demo/ebook/the-lighthouse-keeper.epub kind = compressed_blob sizeBytes = 33071 magicHex = [redacted]

Medium

Ships Compressed Blob

Package ships compressed or archive-like blobs.

demo/ebook/the-lighthouse-keeper.epubView on unpkg

•path = demo/ebook/the-lighthouse-keeper.epub kind = nested_archive_needs_inspection sizeBytes = 33071 magicHex = [redacted]

Low

Nested Archive Needs Inspection

Package ships a nested archive or MCP bundle that was inventoried but not recursively analyzed.

demo/ebook/the-lighthouse-keeper.epubView on unpkg

app/assets/WorkerParser.jsView file

•stage = ast_semantic_analysis; reason = ast_trace_path_limit_exceeded; limitedFiles = 5

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

app/assets/WorkerParser.jsView on unpkg

app/assets/browser.jsView file

•matchType = normalized_sha256 matchedPackage = @knockdata/objectexplorer@0.7.0 matchedPath = app/assets/browser.js matchedIdentity = npm:[redacted]:0.7.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

app/assets/browser.jsView on unpkg

app/assets/wasm-worker.jsView file

•matchType = normalized_sha256 matchedPackage = @knockdata/objectexplorer@0.7.0 matchedPath = app/assets/wasm-worker.js matchedIdentity = npm:[redacted]:0.7.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

app/assets/wasm-worker.jsView on unpkg

app/assets/Workbook.jsView file

•matchType = normalized_sha256 matchedPackage = @knockdata/objectexplorer@0.7.0 matchedPath = app/assets/Workbook.js matchedIdentity = npm:[redacted]:0.7.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

app/assets/Workbook.jsView on unpkg

Source & flagged code

27 flagged · loading source

server/WebServer.mjsView file

Critical

Same File Env Network Execution

A single source file combines environment access, network access, and code or shell execution with blocking evidence.

server/WebServer.mjsView on unpkg · L181

Findings

6 Critical16 High7 Medium6 Low

CriticalSame File Env Network Executionserver/WebServer.mjs

CriticalHardcoded Runtime Data Exfiltrationserver/WebServer.mjs

CriticalCredential Exfiltrationserver/WebServer.mjs

CriticalBrowser Credential Phishingserver/WebServer.mjs

CriticalTrojan Source Unicodeapp/assets/index.js

CriticalTrigger Reachable Dangerous Capabilityserver/WebServer.mjs

HighChild Processserver/WebServer.mjs

HighShellserver/WebServer.mjs

HighEvalapp/assets/index.js

HighCredential Redirect Persistenceserver/WebServer.mjs

HighCommand Output Exfiltrationserver/WebServer.mjs

High

Affected versions and remediation

This report applies to @knockdata/objectexplorer@0.8.0.

See all versions of @knockdata/objectexplorer for other recorded verdicts.

Review the evidence and your use of @knockdata/objectexplorer@0.8.0 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: 2026-09-28 21:40:42.291Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Runtime surface

Behavioral surface

Source & flagged code