Back to skill

Security audit

Eden AI

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly coherent for using Eden AI through OOMOL, but its setup instructions tell users to execute remote installer scripts without verification.

Review the setup instructions before installing. Prefer official package-manager or download-and-verify installation steps for the oo CLI instead of piping remote scripts directly into a shell, and confirm any Eden AI write payload before running 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:56
Finding
Unverified Remote Installer Download and Immediate Execution## Vulnerability Details **File Location**: `SKILL.md`, lines 56–60 **Vulnerability Type**: Remote Payload Retrieval and Execution **Risk Level**: High ### Vulnerable Code ```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 The first-time setup instructions retrieve mutable scripts from an external server and execute them immediately using Bash or PowerShell. They do not pin an installer version, validate a cryptographic checksum or signature, or provide an opportunity to inspect the downloaded content before execution. Installing the `oo` CLI is related to the Skill's declared purpose, and the scripts are hosted on the documented OOMOL domain. However, piping remotely retrieved content directly into a command interpreter is not the minimum safe privilege or trust model necessary to install a client. The effective installer payload is controlled outside the audited package and can change after review. The audit cannot establish that the current remote scripts are malicious because their contents were not included in the project. The confirmed vulnerability is the unsafe remote execution mechanism and its supply-chain exposure. ### Attack Path 1. The `oo` CLI is absent, causing an `oo: command not found` error. 2. The user or Agent follows the first-time setup instructions. 3. `curl` or `Invoke-RestMethod` retrieves the current installer from `cli.oomol.com`. 4. Bash or `Invoke-Expression` executes the response without integrity verification. 5. If the hosting infrastructure, publication pipeline, domain, TLS trust path, or installer itself has been compromised, attacker-controlled commands execute with the invoking user's privileges. ### Impact Assessment A substituted installer could run arbitrary commands as the current user. Depending on that user's permissions ...[truncated 360 chars]
Remediation
## Remediation Suggestions - Replace direct `curl | bash` and `irm | iex` execution with a documented download-and-verify workflow. - Distribute a version-pinned release artifact through a trusted package manager or immutable release URL. - Publish a cryptographic checksum and preferably a verifiable signature for each installer artifact. - Download the installer to a local file, verify its expected origin, hash, and signature, and only then execute it. - Avoid requiring administrator privileges unless a specific installation operation needs them; clearly document any such operation. - Pin the installed CLI version and document a controlled update process so installer behavior cannot change silently after the Skill is audited. - Advise users to inspect the installer and run it in a restricted environment where practical.
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
Findings (2)

External Script Fetching

High
Category
Supply Chain
Content
- **`oo: command not found`** — install the oo CLI (other platforms: <https://cli.oomol.com/install-guide.md>):

  ```bash
  curl -fsSL https://cli.oomol.com/install.sh | bash    # macOS / Linux
  ```

  ```powershell
Confidence
98% confidence
Finding
The skill instructs users to install software via a remote script piped directly into a shell (`curl ... | bash`), which executes unverified code from the network without prior inspection. If the install host, transport, or script supply chain is compromised, this can lead to arbitrary code execution on the user's machine; in a skill context, this is more dangerous because it operationalizes the command as a troubleshooting step users may copy and run verbatim.

Vague Triggers

Medium
Confidence
97% confidence
Finding
The description says to use this skill for "ANY Eden AI request" and "Whenever a task involves Eden AI," which is extremely broad and lacks boundaries or exclusion conditions. For a manifest/markdown-scoped vague-trigger review, this can overlap with many ordinary requests mentioning Eden AI without clarifying when the skill should not activate.

Static analysis

No suspicious patterns detected.