Back to skill

Security audit

Dice Easy Apply Automation

Security checks for vulnerabilities and agentic risk

Overview

The skill matches its Dice job-application purpose, but its default browser setup disables Chromium sandboxing while using a persistent logged-in profile and uploading resume documents.

Review before installing or running. Use a dedicated low-privilege account and isolated browser profile, remove the sandbox-disabling Chromium flags where possible, keep DRY_RUN enabled until you verify behavior, and only point RESUME_PDF/COVER_PDF at documents you intend to submit to Dice applications.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:99
Finding

Chromium Security Sandbox Explicitly Disabled

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 99-105
Vulnerability Type: Browser sandbox disabled through insecure launch flags
Risk Level: Medium

js
const browser = await puppeteer.launch({
  headless: false,
  executablePath: process.env.CHROME_BIN || '/snap/bin/chromium',
  userDataDir: process.env.CHROME_PROFILE || `${process.env.HOME}/.cache/dice-chrome`,
  defaultViewport: null,
  args: ['--no-sandbox', '--disable-setuid-sandbox', '--disable-dev-shm-usage', '--start-maximized']
});

Technical Analysis

The recommended Puppeteer configuration passes both --no-sandbox and --disable-setuid-sandbox to Chromium. These flags disable important operating-system isolation boundaries intended to contain a compromised browser renderer.

This workflow visits remote web content while using a persistent, authenticated Chromium profile and uploading sensitive employment documents. If malicious page content or a compromised third-party resource exploits a Chromium vulnerability, disabling the sandbox can make it easier for the exploit to reach the privileges and local resources available to the browser process. The configuration is therefore unsafe for routine execution on a workstation or host containing sensitive information.

Attack Path

  1. An operator launches the documented automation with the sandbox-disabling flags.
  2. Chromium loads Dice pages and remotely hosted browser resources while using a persistent authenticated profile.
  3. Malicious or compromised web content exploits a compatible Chromium vulnerability.
  4. Because the browser sandbox is disabled, the exploit has a reduced isolation barrier between renderer content and the agent user's environment.
  5. Successful host-level execution could access resources available to the browser process, including local files permitted to that user, uploaded documents, browser-profile data, and application state.

This path requires a ...[truncated 671 chars]

Remediation
View remediation

Remediation Suggestions

  • Remove --no-sandbox and --disable-setuid-sandbox from the Chromium arguments and retain the browser's supported sandbox configuration.
  • Run Chromium as a dedicated, unprivileged operating-system user rather than as root or an administrative account.
  • Keep Chromium and Puppeteer patched to supported versions so known browser vulnerabilities are addressed promptly.
  • Store the persistent profile with restrictive filesystem permissions and use it only for this workflow.
  • Limit the automation account's access to the specific resume, cover letter, state directory, and required browser files.
  • If sandbox-disabling flags are unavoidable in a constrained environment, run the browser in a disposable, strongly isolated container or virtual machine with no sensitive host mounts, no unnecessary credentials, restricted networking, dropped Linux capabilities, and a read-only filesystem wherever practical.
  • Document the sandbox requirement as an exceptional compatibility fallback rather than presenting it as the default launch pattern.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep