Back to skill

Security audit

HaiGuanJia

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent HaiGuanJia connector, but its setup instructions include direct remote installer execution and it can reach write, delete, and billed business actions.

Install only if you trust OOMOL and the HaiGuanJia connector workflow. Before running setup, prefer a verified package or reviewed installer, avoid elevated shells, and carefully confirm any write, delete, subscription, or chargeable action payload before approving it.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:126
Finding
Unverified Remote Installation Scripts Are Executed Directly by the Shell## Vulnerability Details **File Location**: `SKILL.md`, lines 126–134 **Vulnerability Type**: `T03: Remote Payload Retrieval and Execution` **Risk Level**: High The first-time setup instructions recommend downloading mutable scripts from an external server and immediately executing them: ```bash curl -fsSL https://cli.oomol.com/install.sh | bash # macOS / Linux ``` ```powershell irm https://cli.oomol.com/install.ps1 | iex # Windows PowerShell ``` ### Technical Analysis Both installation methods connect a network retrieval command directly to a command interpreter. Neither method pins a specific release nor verifies a cryptographic signature or checksum before execution. Consequently, the code ultimately executed can change after the Skill has been reviewed. Installing the `oo` CLI is relevant to the Skill's declared connector functionality, and the referenced domain appears consistent with the declared OOMOL service. However, direct download-to-shell execution is not the minimum safe installation mechanism. It grants the remote script the full privileges of the user running the agent and provides no meaningful opportunity to inspect the payload before execution. The same weakness exists on both supported platform families: - On macOS and Linux, `curl` sends the downloaded response directly to `bash`. - On Windows, `Invoke-RestMethod` downloads the response and `Invoke-Expression` evaluates it as PowerShell code. No evidence in the audited file establishes that the remote scripts are currently malicious. The vulnerability is that compromise of the delivery path or vendor infrastructure would turn these documented setup commands into an arbitrary code-execution channel. ### Attack Path 1. The `oo` command is unavailable, causing the agent or user to consult the first-time setup instructions. 2. An attacker compromises the installation server, its deployment pipeline, or another trusted component in th ...[truncated 1529 chars]
Remediation
## Remediation Suggestions 1. Remove the `curl | bash` and `irm | iex` installation patterns. 2. Direct users to a documented package manager or a versioned release artifact from the official distribution channel. 3. Pin the CLI to a specific reviewed version rather than retrieving an unversioned installation script. 4. Publish a cryptographic checksum or signature through an independently protected release channel. 5. Require users to download the artifact first, verify its checksum or signature, and only then execute or install it. 6. Present the exact artifact URL, expected version, expected digest, and verification commands in the setup instructions. 7. Require explicit user approval before installing software; the agent should not perform installation automatically after a command failure. 8. Recommend installation from a non-elevated account and clearly warn users not to run the installer with administrative or root privileges unless strictly required. 9. If an installation script remains necessary, pin its immutable revision and verify the downloaded content before invoking the interpreter. For example, download to a file, validate a published SHA-256 digest or signature, allow inspection, and execute it as a separate step. 10. Protect release infrastructure with signed releases, restricted deployment permissions, multifactor authentication, and auditable build provenance.
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (1)

Vague Triggers

Medium
Confidence
94% confidence
Finding
The skill description claims it should be used for ANY HaiGuanJia request, including read, create, update, and delete operations. This overly broad trigger can cause the agent to route all HaiGuanJia-related tasks through a high-privilege skill, increasing the chance of unintended writes, destructive actions, or billed operations being selected when a narrower or safer workflow would be more appropriate.

Static analysis

No suspicious patterns detected.