Back to skill

Security audit

boss-job

Security checks for vulnerabilities and agentic risk

Overview

This is a plausible BOSS job-search skill, but it asks users to install mutable external components that operate through a logged-in Chrome session and can read or send recruiter messages.

Install only if you trust the OpenCLI package, the GitHub plugin repository, and the Chrome extension source. Before using greet or send, verify the recipient and message text yourself, and avoid letting an agent send messages automatically or in bulk from your logged-in BOSS直聘 account.

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

Warning
Location
SKILL.md:14
Finding

Unpinned Third-Party Components Installed from Mutable Remote Sources

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 14–22
Vulnerability Type: T08: Insecure Dependencies
Risk Level: Medium

The skill directs users to globally install an npm package, install an unspecified Chrome extension, and install a plugin directly from a mutable GitHub repository:

bash
npm install -g @jackwener/opencli
bash
opencli plugin install github:SPYQWER1/opencli-plugin-boss-job

Technical Analysis

Neither installation command pins the dependency to a reviewed version or immutable commit. No integrity hash, checksum, or signature-verification procedure is provided. The source and version of the required Chrome extension are also unspecified.

Installing the npm package globally may expose system-level CLI resources available to the current user. The GitHub plugin reference can resolve to different content over time without any modification to this skill. Because the documented workflow reuses an authenticated Chrome session for zhipin.com, the installed components may operate in a sensitive context containing session cookies, job-search information, recruiter conversations, and message-sending capabilities.

The repository contains only SKILL.md; therefore, there is no bundled executable code demonstrating an intentional malicious payload. The risk arises from the unpinned and insufficiently verified dependency-installation process.

Attack Path

  1. An attacker compromises the npm publisher account, GitHub repository, release process, or distribution channel for the Chrome extension.
  2. The attacker publishes malicious code under the same package, repository reference, or extension identity.
  3. A user follows the skill instructions and installs the current mutable dependency without version or integrity verification.
  4. The compromised component executes in the user's environment and interacts with OpenCLI or the browser extension.
  5. When the user is logged i ...[truncated 987 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin @jackwener/opencli to a specifically reviewed version rather than installing the latest published release.
  2. Record and verify npm integrity metadata or an independently published cryptographic checksum before installation.
  3. Pin the GitHub plugin to a reviewed immutable commit hash or signed release instead of a mutable repository reference.
  4. Document the official Chrome extension source, publisher identity, exact version, required permissions, and signature-verification procedure.
  5. Review the source code of OpenCLI, the plugin, and the extension together because the security boundary spans all three components.
  6. Require explicit user confirmation before reading conversations, greeting recruiters, or sending messages.
  7. Grant the extension only the minimum host permissions needed for zhipin.com, and avoid access to unrelated sites or browser data.
  8. Run the CLI with an unprivileged user account and avoid elevated installation or execution where possible.
  9. Maintain a dependency lock or manifest containing approved versions, commit hashes, checksums, and review dates.
  10. Re-review dependencies before updating any pinned version or commit.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly supports sending greetings and chat messages through the user's existing logged-in BOSS直聘 session but does not warn that these commands will perform real external actions on the user's account and transmit user-provided text to third parties. This creates a meaningful risk of unintended outbound communication, account misuse, and privacy leakage, especially if an agent invokes the skill without clear confirmation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The workflow examples normalize automated recruiter outreach and replies without any safety notice that these are real-world account actions affecting external parties. In an agent setting, such examples increase the chance that messaging is executed implicitly or at scale, causing unintended contact, reputational harm, or disclosure of personal information from the user's account.

Content

No source excerpt is available for this finding.