Back to skill

Security audit

浏览器账号控制器

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed local controller for a browser account management tool, with setup supply-chain cautions but no evidence of hidden or malicious behavior.

Install only if you trust the third-party browser account management application and are comfortable letting a local API start, stop, and list those browser accounts. Prefer verifying the external installer and using a controlled Python environment before running pip install requests.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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)

T08 · Insecure Dependencies

Note
Location
SKILL.md:82
Finding

Unpinned Dependency and Unverifiable External Installer Provenance

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 82–87
Vulnerability Type: Supply-chain risk from an unpinned Python dependency and an externally distributed application without documented integrity verification
Risk Level: Suspicious

Vulnerable Code Snippet

bash
pip install requests
text
The “multi-account matrix management tool” must be installed from:
https://zmt.scys6688.com/

Technical Analysis

The setup instructions install requests without specifying an exact version, dependency lockfile, or cryptographic hashes. Consequently, the package and its transitive dependency versions can change after the Skill has been reviewed.

The instructions also direct users to obtain a browser-account management application from an external vendor website without documenting an exact application version, SHA-256 checksum, expected publisher certificate, or signature-verification procedure. The external application is not included in this repository and therefore was outside the inspected code boundary.

This creates a supply-chain trust gap. A compromised package index, dependency release, vendor website, download channel, or vendor installer could cause users to install code that differs from the components originally expected by the Skill.

Attack Path

  1. An attacker compromises a relevant package release, package distribution channel, vendor website, or downloadable installer.
  2. The user follows the Skill instructions and runs the unpinned installation command or downloads the external application.
  3. The substituted component executes with the privileges of the installing or launching user.
  4. The malicious component can access resources available to that user.
  5. Because the external application manages browser accounts and exposes the local API used by the Skill, a compromised component may manipulate account operations or attempt to access associated browser-session information.

No compromise of either distri ...[truncated 679 chars]

Remediation
View remediation

Remediation Suggestions

  1. Pin requests and all transitive dependencies to reviewed versions in a lockfile.
  2. Record cryptographic hashes for every Python distribution and install with hash enforcement, for example:
    bash
    pip install --require-hashes -r requirements.txt
    
  3. Use a controlled or trusted package index and document the expected index URL.
  4. Specify the exact supported version and canonical HTTPS download URL for the external account-management application.
  5. Publish and verify a SHA-256 checksum for the downloaded installer.
  6. Document the expected publisher identity and require verification of the installer’s operating-system code signature before execution.
  7. Update the documented version, checksum, and signature information through a reviewed release process.
  8. Run the external application with the minimum necessary user privileges and limit access to browser data unrelated to its intended account-management function.
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