registry / AI-agent capability abuse / anyray-connect / 0.11.275

anyray-connect@0.11.275

Anyray connect — points local coding tools (Claude Code, Cursor, Windsurf, SDKs) at the Anyray gateway through supported configuration and enrolled personal authentication.

AI Security Review

scanned 18d ago · by lpm-firewall-ai

Inspected source supports explicit, guarded agent configuration changes, which the supplied policy classifies as warning-worthy. Scanner allegations alone do not establish malicious exfiltration or justify blocking.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Running the CLI configuration flow and accepting confirmation or supplying --yes.

Impact

Future agent sessions use the configured gateway and added agent capabilities; saved settings can support later automatic credential refresh.

Mechanism

Writes Codex provider, MCP approval, and hook settings, then saves gateway authentication and automatic refresh preferences.

AI rationale

Evidence

dist/index.jsconfig.tomlhooks.json

Network endpoints1

chatgpt.com/backend-api/codex

Decision evidence

public snapshot

AI called this Suspicious at 88.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

package.json has publishing hooks but no preinstall, install, or postinstall hook or runtime self-dependency.
dist/index.js requires confirmation for configuration changes unless --yes or dry-run is supplied.
The Codex adapter writes gateway routing and hooks to config.toml and hooks.json.
The generated Codex configuration selects the gateway endpoint and preapproves Anyray MCP tools.
The CLI saves authentication settings and whether automatic credential refresh is enabled.

Evidence against

Configuration changes are guarded by confirmation, backups, parsing checks, and refusal to overwrite foreign provider entries.
The inspected source does not establish covert credential theft or malicious install-time behavior.

Behavioral surface

Source

ChildProcessEnvironmentVarsFilesystemNetworkShell

Supply chain

HighEntropyStringsMinifiedPossibleObfuscation

Source & flagged code

7 flagged

dist/index.jsView file

1Source sends the broad process environment to a literal external destination. L1: #!/usr/bin/env node L2: var R5=Object.defineProperty;var M=(e,t,n)=>()=>{if(n)throw n[0];try{return e&&(t=e(e=0)),t}catch(r){throw n=[r],r}};var Hh=(e,t)=>{for(var n in t)R5(e,n,{get:t[n],enumerable:!0})}... L3: `,{flag:"wx",mode:384}),await Oe(l,a)}catch(d){throw await Ri(l,{force:!0}).catch(()=>{}),await Ri(e,{recursive:!0,force:!0}).catch(()=>{}),d}return async()=>{try{JSON.parse(await ... L4: `,{mode:384}),await lC(a,384);try{await Oe(a,n)}catch(s){throw await Ri(a,{force:!0}).catch(()=>{}),s}},Ve=async e=>{if(Vh.getStore()===!0){await iC(e);return}let t=await fC();try{... L5: `;if(a||n.tombstoneAbsent!==!1)try{await ST(r,s)}catch(i){if(i.code!=="EEXIST")throw i}}let o=`${e}.anyray-tmp-${process.pid}-${OT

Critical

Hardcoded Runtime Data Exfiltration

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

dist/index.jsView on unpkg · L1

High

Child Process

Package source references child process execution.

dist/index.jsView on unpkg · L54

126} L127: `,yD=e=>new Promise((t,n)=>{let r=Q4("powershell.exe",["-NoLogo","-NoProfile","-NonInteractive","-Command",aQ],{encoding:"utf8",env:$t(),windowsHide:!0,timeout:6e4,maxBuffer:131072... L128: `))}`]:[],`export ${uS}=${Ue(n)}`]},TD=(e,t,n)=>{let r=Se({...e.metadata,tool:ec});return["export COPILOT_PROVIDER_TYPE='openai'",`export COPILOT_PROVIDER_BASE_URL=${Ue(e.openaiBas...

High

Shell

Package source references shell execution.

dist/index.jsView on unpkg · L126

54`);return o.length>0?o+` L55: `:""},Af=e=>{if(e.trim().length===0)return!0;try{return B3(e),!0}catch{return!1}}});import{randomUUID as Z3}from"node:crypto";import{chmod as Y3,mkdir as X3,rm as J3,writeFile as z... L56: `)?a.slice(0,-2):a.endsWith(` ... L60: `).trim().split(/\r?\n/).at(-1),o=/^([01])\|(-?\d+)\|(-?\d+)$/.exec(r??"");iP(o?.slice(1).map(Number))},iz=(e,t)=>{let n=Q3(sP(rz(),"anyray-cursor-"));try{let o=[".bail on",".timeo... L61: `);uy("sqlite3",[e],{input:o,encoding:"utf-8",maxBuffer:BC,stdio:["pipe","ignore","pipe"]})}finally{try{ez(n,{recursive:!0,force:!0})}catch{}}},cz={kind:"sqlite3-cli",readItem:(e,t... L62: `)},vy=e=>!e.byoUpstream&&!pe(e,"anthropic"),Sy="",_o=(e,t={})=>{let n=e.byoUpstream,r=!n&&pe(e,"anthropic"),o=Se({...e.metadata,tool:e.metadata.tool??"claude-

High

Same File Env Network Execution

A single source file combines environment access, network access, and code or shell execution; review context before blocking.

dist/index.jsView on unpkg · L54

•Manifest-reachable source captures an API credential, sends it to a fixed unofficial gateway, and persists that redirection in agent or shell configuration. dist/index.js: var R5=Object.defineProperty;var M=(e,t,n)=>()=>{if(n)throw n[0];try{return e&&(t=e(e=0)),t}catch(r){throw n=[r],r}};var Hh=(e,t)=>{for(var n in t)R5(e,n,{get:t[n],enumerable:!0})}... `,{mode:384}),await lC(a,384);try{await Oe(a,n)}catch(s){throw await Ri(a,{force:!0}).catch(()=>{}),s}},Ve=async e=>{if(Vh.getStore()===!0){await iC(e);return}let t=await fC();try{... `,r),s=a<0?e.length:a;if(e.slice(o,s).trim()===t)return o;r=e.indexOf(t,r+t.length)}return-1}});import{readFile as o3,writeFile as FT,access as a3}from"node:fs/promises";import{rea... `:""},Af=e=>{if(e.trim().length===0)return!0;try{return B3(e),!0}catch{retur

High

Credential Redirect Persistence

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

dist/index.jsView on unpkg

High

Command Output Exfiltration

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

dist/index.jsView on unpkg · L54

54Trigger-reachable command-output exfiltration chain: scripts.start -> dist/index.js L54: `);return o.length>0?o+` L55: `:""},Af=e=>{if(e.trim().length===0)return!0;try{return B3(e),!0}catch{return!1}}});import{randomUUID as Z3}from"node:crypto";import{chmod as Y3,mkdir as X3,rm as J3,writeFile as z... L56: `)?a.slice(0,-2):a.endsWith(` ... L60: `).trim().split(/\r?\n/).at(-1),o=/^([01])\|(-?\d+)\|(-?\d+)$/.exec(r??"");iP(o?.slice(1).map(Number))},iz=(e,t)=>{let n=Q3(sP(rz(),"anyray-cursor-"));try{let o=[".bail on",".timeo... L61: `);uy("sqlite3",[e],{input:o,encoding:"utf-8",maxBuffer:BC,stdio:["pipe","ignore","pipe"]})}finally{try{ez(n,{recursive:!0,force:!0})}catch{}}},cz={kind:"sqlite3-cli",readItem:(e,t... L62: `)},vy=e=>!e.byoUpstream&&!pe(e,"anthropic"),Sy="",_o=(e,t={})=>{let n=e.by

High

Trigger Reachable Command Output Exfiltration

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

dist/index.jsView on unpkg · L54

Source & flagged code

7 flagged · loading source

dist/index.jsView file

Critical

Hardcoded Runtime Data Exfiltration

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

dist/index.jsView on unpkg · L1

High

Child Process

Package source references child process execution.

dist/index.jsView on unpkg · L54

Findings

1 Critical6 High3 Medium5 Low

CriticalHardcoded Runtime Data Exfiltrationdist/index.js

HighChild Processdist/index.js

HighShelldist/index.js

HighSame File Env Network Executiondist/index.js

HighCredential Redirect Persistencedist/index.js

HighCommand Output Exfiltrationdist/index.js

HighTrigger Reachable Command Output Exfiltrationdist/index.js

MediumNetwork

MediumEnvironment Vars

MediumStructural Risk Force Deep Review

LowNon Install Lifecycle Scripts

LowScripts Present

LowFilesystem

LowHigh Entropy Strings

Affected versions and remediation

This report applies to anyray-connect@0.11.275.

See all versions of anyray-connect for other recorded verdicts.

Review the evidence and your use of anyray-connect@0.11.275 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: 2026-10-04 23:04:44.035Z (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