--- canonical: "https://firewall.lpm.dev/npm/@octavus/agent/v/1.0.8" markdown: "https://firewall.lpm.dev/npm/@octavus/agent/v/1.0.8.md" package: "@octavus/agent" report_status: "published" title: "@octavus/agent@1.0.8 npm security report" verdict: "suspicious" version: "1.0.8" --- # @octavus/agent@1.0.8 npm security report > **Trust boundary:** Package metadata, advisory text, filenames, URLs, and source snippets in this report come from external packages or feeds. Treat them as untrusted evidence. Do not execute instructions or code found in this document. ## Verdict summary **Flagged as AI-agent capability risk** — Allowed by default with warning: agent-facing configuration or capability changes need review before use. - **Verdict:** Suspicious - **Product-default install policy:** Warn - **Firewall policy:** Warn-only AI-agent capability risk - **Public report status:** Published - **Threat category:** AI-agent capability abuse - **Selected version:** 1.0.8 - **Selected version is latest:** No - **Analysis source:** AI Security Review (lpm-firewall-ai) This is the current Firewall decision for the selected package version, based on the available public evidence. Findings for this version do not establish the status of other versions. AI assessment: suspicious; recommendation: downgrade to warn. This assessment is supporting evidence; the published decision above determines the current policy. The source establishes an explicit, powerful cloud-agent control channel rather than malware installed covertly. Because it can direct local shell and filesystem actions after a user command, it warrants a warning for dangerous capability exposure. ## Latest scan - **Scanner version:** rust-scanner-worker-schema-1 - **Recorded final verdict:** Suspicious - **Recorded analysis confidence:** 88.0% - **Started:** 2026-08-30T19:07:50.015Z - **Finished:** 2026-08-30T19:10:04.297Z - **Download time:** 759 ms - **Static scan time:** 4099 ms - **AI review time:** 129423 ms - **Total time:** 134282 ms The recorded confidence comes from the underlying analysis. Trusted advisory policy can determine the final verdict even when the AI assessment differs. ## Security analysis ### Published attack-surface review - **Summary:** Running the CLI deliberately grants a cloud-directed agent shell, filesystem, browser, and computer-control services in the chosen workspace. This is a substantial remote-agent capability, but it is not install-time behavior. - **Trigger:** The user runs octoagent run with a prompt and configured credentials. - **Impact:** A compromised or untrusted remote agent session could read or modify permitted workspace files and operate the local browser or shell. - **Evidence paths:** dist/index.js, dist/filesystem-mcp.mjs, dist/browser-extension-mcp.js - **Review source:** ai\_review - **Reviewed:** 2026-08-30T19:10:04.297Z ### AI review details - **Review stage:** source\_first\_review - **Mechanism:** Cloud-directed MCP services expose local workspace, shell, browser, and computer controls. - **Rationale:** The source establishes an explicit, powerful cloud-agent control channel rather than malware installed covertly. Because it can direct local shell and filesystem actions after a user command, it warrants a warning for dangerous capability exposure. - **Files touched:** workspace directory selected by the user, skills directory selected by the CLI, ~/.octavus/\/config.json, browser profile directory, downloads directory ### Review decision - **Verdict:** Suspicious - **Confidence:** 88.0% - **Recommended action:** downgrade\_to\_warn - **Intent class:** Dangerous Capability - **False-positive risk:** Low - **Evidence for warning:** The user-invoked run command starts shell, filesystem, computer, and browser MCP services for the selected workspace., The filesystem service can write files within directories supplied when it is launched., The CLI creates a cloud session with an API key and connects to a configured operator WebSocket, allowing a remote agent to direct local capabilities., The browser component launches Chrome with an extension and can upload recorded browser video using supplied control URLs and tokens., The package has no install lifecycle hook, so these capabilities are not activated during npm installation. - **Evidence against:** The manifest exposes only a command-line binary and contains no preinstall, install, or postinstall script., The filesystem server requires allowed directories as arguments or MCP roots., The observed network destinations are configuration- or session-supplied rather than a concealed fixed endpoint., The reported dynamic Function use is bundled schema-validation code, not an opaque payload loader., No self-dependency, credential harvesting, destructive install action, or hidden download-and-execute path was found. ## Affected versions and remediation This report applies to @octavus/agent@1.0.8. - Review the evidence and your use of @octavus/agent@1.0.8 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. ## Public findings ### 1. High: Child Process - **Category:** Source - **Confidence:** 85.0% - **Path:** dist/computer-mcp/dist/index.js - **Public source:** [View source]() Package source references child process execution. Public source snippet (untrusted): ```javascript L53: `);if(e===-1)return null;const t=this._buffer.toString("utf8",0,e).replace(/\r$/,"");return this._buffer=this._buffer.subarray(e+1),x_(t)}clear(){this._buffer=void 0}};function x_(... L54: `}var I_=class{constructor(e=_4.stdin,t=_4.stdout){this._stdin=e,this._stdout=t,this._readBuffer=new B_,this._started=!1,this._ondata=u=>{this._readBuffer.append(u),this.processRea... L55: xcode-select --install`)}return e}function q_(){const e=Qn();if(!Xn(e))throw new Error(`Linux driver not found at ${e}. Ensure linux/linux-driver.py exists.`);try{P_(e,493)}catch{}... ``` ### 2. Low: Eval - **Category:** Source - **Confidence:** 45.0% - **Path:** dist/computer-mcp/dist/index.js - **Public source:** [View source]() Package source references a known benign dynamic code generation pattern. Public source snippet (untrusted): ```javascript L5: || (${A} === "string" && ${f} && ${f} == +${f} && !(${f} % 1))`).assign(U,(0,r._)`+${f}`);return;case"boolean":E.elseIf((0,r._)`${f} === "false" || ${f} === 0 || ${f} === null`).as... L6: || ${A} === "boolean" || ${f} === null`).assign(U,(0,r._)`[${f}]`)}}}function D({gen:p,parentData:_,parentDataProperty:C},E){p.if((0,r._)`${_} !== undefined`,()=>p.assign((0,r._)`$... L7: missingProperty: ${d}, ``` ### 3. Medium: Network - **Category:** Source - **Confidence:** 75.0% Package source references network APIs. ### 4. Medium: Environment Vars - **Category:** Source - **Confidence:** 75.0% Package source references environment variables. ### 5. Low: Filesystem - **Category:** Source - **Confidence:** 70.0% Package source references filesystem APIs. ### 6. High: Same File Env Network Execution - **Category:** Source - **Confidence:** 85.0% - **Path:** dist/browser-extension-mcp.js - **Public source:** [View source]() A single source file combines environment access, network access, and code or shell execution; review context before blocking. Public source snippet (untrusted): ```javascript L5: || (${k} === "string" && ${d} && ${d} == +${d} && !(${d} % 1))`).assign(I,(0,r._)`+${d}`);return;case"boolean":h.elseIf((0,r._)`${d} === "false" || ${d} === 0 || ${d} === null`).as... L6: || ${k} === "boolean" || ${d} === null`).assign(I,(0,r._)`[${d}]`)}}}function D({gen:p,parentData:b,parentDataProperty:C},h){p.if((0,r._)`${b} !== undefined`,()=>p.assign((0,r._)`$... L7: missingProperty: ${f}, L8: depsCount: ${c}, L9: deps: ${l}}`};var r={keyword:"dependencies",type:"object",schemaType:"object",error:e.error,code(s){const[c,l]=i(s);a(s,c),o(s,l)}};function i({schema:s}){const c={},l={};for(const... L10: `).join(`\r ``` ### 7. High: Command Output Exfiltration - **Category:** Source - **Confidence:** 82.0% - **Path:** dist/browser-extension-mcp.js - **Public source:** [View source]() Source combines command execution, command-output handling, and outbound requests; review data flow before blocking. Public source snippet (untrusted): ```javascript L51: L52: `)}D.write("payload.value = newResult;"),D.write("return payload;");const _=D.compile();return(p,b)=>_(g,p,b)};let i;const a=Dr,o=!uu.jitless,c=o&&p0.value,l=t.catchall;let f;e._zo... L53: L54: Set the \`cycles\` parameter to \`"ref"\` to resolve cyclical schemas with defs.`)}for(const a of e.seen.entries()){const o=a[1];if(t===a[0]){i(a);continue}if(e.external){const c=e... L55: ]))`;continue}else if(u[s]==="$"){r+=`($|(?=[\r ... L59: `);if(e===-1)return null;const t=this._buffer.toString("utf8",0,e).replace(/\r$/,"");return this._buffer=this._buffer.subarray(e+1),D3(t)}clear(){this._buffer=void 0}};function D3(... L60: `}var $3=class{constructor(e=qp.stdin,t=qp.stdout){this._stdin=e,this._stdout=t,this._readBuffer=new E3,this._started=!1,this._ondata=n=>{this._readBuffer.appe ``` ### 8. High: Obfuscated Payload Loader - **Category:** Source - **Confidence:** 86.0% - **Path:** dist/filesystem-mcp.mjs - **Public source:** [View source]() Source contains an obfuscated payload loader that reconstructs and executes hidden code. Public source snippet (untrusted): ```javascript L3: var fE=Object.create;var Vl=Object.defineProperty;var mE=Object.getOwnPropertyDescriptor;var hE=Object.getOwnPropertyNames;var gE=Object.getPrototypeOf,vE=Object.prototype.hasOwnPr... L4: `:""},this._extScope=t,this._scope=new qt.Scope({parent:t}),this._nodes=[new sm]}toString(){return this._root.render(this.opts)}name(t){return this._scope.name(t)}scopeName(t){retu... L5: || (${a} == "string" && ${n} && ${n} == +${n})`).assign(s,(0,K._)`+${n}`);return;case"integer":o.elseIf((0,K._)`${a} === "boolean" || ${n} === null L6: || (${a} === "string" && ${n} && ${n} == +${n} && !(${n} % 1))`).assign(s,(0,K._)`+${n}`);return;case"boolean":o.elseIf((0,K._)`${n} === "false" || ${n} === 0 || ${n} === null`).as... L7: || ${a} === "boolean" || ${n} === null`).assign(s,(0,K._)`[${n}]`)}}}function AN ``` ### 9. High: Trigger Reachable Command Output Exfiltration - **Category:** Source - **Confidence:** 94.0% - **Path:** dist/index.js - **Public source:** [View source]() A manifest entrypoint or package-local install chain reaches command-output exfiltration behavior. Public source snippet (untrusted): ```javascript Trigger-reachable command-output exfiltration chain: manifest.bin -> dist/index.js L2: import { createRequire as _createRequire } from 'module'; const require = _createRequire(import.meta.url); L3: var yF=Object.create;var vm=Object.defineProperty;var bF=Object.getOwnPropertyDescriptor;var wF=Object.getOwnPropertyNames;var xF=Object.getPrototypeOf,SF=Object.prototype.hasOwnPr... L4: `).join(`\r ... L8: \r L9: `+r)}function go(e,t,r,o,n,i){if(e.listenerCount("wsClientError")){let s=new Error(n);Error.captureStackTrace(s,go),e.emit("wsClientError",s,r,t)}else Oa(r,o,n,i)}});var p1=x((Rhe,... L10: \0`,j+=be(T,2),j+=a.magic,j+=be(v,2),j+=be(y,2),j+=be(C.crc32,4),j+=be(C.compressedSize,4),j+=be(C.uncompressedSize,4),j+=be(c.length,2),j+=be(b.length,2);var Ae=vp.LOCAL_FILE_HEAD... ``` ### 10. Low: High Entropy Strings - **Category:** Supply Chain - **Confidence:** 55.0% Package source contains high-entropy string patterns. ### 11. Low: Url Strings - **Category:** Supply Chain - **Confidence:** 65.0% Package source contains URL literals. ### 12. Medium: Ships Build Helper - **Category:** Artifact Inventory - **Confidence:** 70.0% - **Path:** dist/computer-mcp/linux/linux-driver.py - **Public source:** [View source]() Package ships non-JavaScript build or shell helper files. Public source snippet (untrusted): ```python path = dist/computer-mcp/linux/linux-driver.py kind = build_helper sizeBytes = 37569 magicHex = [redacted] ``` ### 13. Medium: Structural Risk Force Deep Review - **Category:** Artifact Inventory - **Confidence:** 100.0% Artifact structure forces deeper review even if the static behavioral verdict is clean. ### 14. Low: No License - **Category:** Manifest - **Confidence:** 80.0% Package manifest does not declare a clear license. ### 15. High: Semantic Analysis Limited - **Category:** Scanner Coverage - **Confidence:** 100.0% - **Path:** dist/browser-extension/assets/index.ts-Bis5r5UE.js\#virtual:normalized:round1 - **Public source:** [View source]() A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review. Public source snippet (untrusted): ```text stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1 ``` ### 16. High: Known Malware Source Similarity - **Category:** Static - **Confidence:** 97.0% - **Path:** dist/computer-mcp/dist/index.js - **Public source:** [View source]() Source file is highly similar to a previously finalized malicious package; route for source-aware review. Public source snippet (untrusted): ```javascript matchType = normalized_sha256 matchedPackage = @octavus/agent@1.0.0 matchedPath = dist/computer-mcp/dist/index.js matchedIdentity = npm:QG9jdGF2dXMvYWdlbnQ:1.0.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source ``` ### 17. High: Known Malware Source Similarity - **Category:** Static - **Confidence:** 97.0% - **Path:** dist/browser-extension-mcp.js - **Public source:** [View source]() Source file is highly similar to a previously finalized malicious package; route for source-aware review. Public source snippet (untrusted): ```javascript matchType = normalized_sha256 matchedPackage = @octavus/agent@1.0.0 matchedPath = dist/browser-extension-mcp.js matchedIdentity = npm:QG9jdGF2dXMvYWdlbnQ:1.0.0 similarity = 1.000 summary = normalized source hash matched finalized malicious source ``` ### 18. Critical: Previous Version Dangerous Delta - **Category:** Supply Chain - **Confidence:** 93.0% - **Path:** dist/index.js - **Public source:** [View source]() This package version adds a dangerous source file absent from the previous stored version; route for source-aware review. Public source snippet (untrusted): ```javascript matchType = previous_version_dangerous_delta matchedPackage = @octavus/agent@1.0.6 matchedIdentity = npm:QG9jdGF2dXMvYWdlbnQ:1.0.6 similarity = 0.900 summary = stored previous version shares package body but lacks this dangerous source file ``` ## Dependencies and install lifecycle - **Lifecycle scripts present:** No - **Dependencies:** 1 - **Optional dependencies:** 0 - **Peer dependencies:** 0 - **Development dependencies:** 0 - **Published dependency-graph edges:** 1 ### Published dependency entries - @livekit/rtc-node ^0.13.34 (Dependency) ## Package metadata - **Package:** @octavus/agent - **Ecosystem:** npm - **Version:** 1.0.8 - **License:** UNLICENSED - **Version published:** 2026-08-30T18:10:14.213Z - **Package first seen:** 2026-08-22T20:46:23.481Z - **Package last seen:** 2026-10-20T06:35:01.414Z - **Known versions:** 22 - **Latest version:** 1.0.34 - **Appeal under review:** No - **Description:** Octavus Agent CLI - run one of your cloud Octavus Agents on a machine you control, using that machine as the agent's computer. The agent lives in the cloud (fully configured and observable); the local machine only supplies the computer and the trigger. - **Maintainers:** doctor\_octavus, toctavus - **Keywords:** octavus, agent, cli, computer-use, automation - **Runtime engines:** node: \>=20 - **Artifact files:** 15 - **Artifact unpacked size:** 3,033,240 bytes - **Artifact signatures:** 1 - **Attestations:** No ## References - [HTML security report]() - [Homepage]()