Back to skill

Security audit

OOMOL Fusion API

Security checks for vulnerabilities and agentic risk

Overview

The skill’s normal API use is mostly coherent, but its setup instructions include unverified remote installer commands that could execute arbitrary local code.

Install only if you trust OOMOL and are comfortable reviewing or replacing the CLI setup steps. Before running first-time setup, prefer a verified package or inspect and verify the installer separately; treat write and destructive API actions as requiring explicit confirmation of the exact target and payload.

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:136
Finding
Unverified Remote Installer Retrieval and Execution## Vulnerability Details **File Location**: `SKILL.md`, lines 136–140 **Vulnerability Type**: Remote payload retrieval and execution **Risk Level**: Critical The first-time setup instructions execute remotely downloaded installer scripts directly in local command interpreters: ```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 commands retrieve mutable content from a remote endpoint and immediately execute it through Bash or PowerShell. The instructions provide no pinned release version, cryptographic signature verification, checksum validation, package-manager verification, or opportunity to inspect the downloaded script before execution. HTTPS protects data in transit but does not establish that the script itself is safe. If the hosting infrastructure, domain, deployment pipeline, or publisher account is compromised—or if the mutable installer is changed after this Skill is reviewed—the effective payload can become malicious without any modification to the audited project. The domains are consistent with the declared OOMOL service, and the audit found no evidence that they are unrelated attacker-controlled hosts. Nevertheless, direct remote-to-shell execution creates an unrestricted code-execution channel and exceeds the minimum privilege necessary to provide installation guidance. ### Attack Path 1. The `oo` command is unavailable, causing the user or Agent to follow the first-time setup instructions. 2. The host requests the current `install.sh` or `install.ps1` content from `cli.oomol.com`. 3. An attacker who has compromised the hosting service, publishing pipeline, domain, or installer replaces the expected script with a malicious payload. 4. The pipe sends the response directly to Bash, or `Invoke-Expression` evaluates the PowerShell response. 5. The malicio ...[truncated 934 chars]
Remediation
## Remediation Suggestions 1. Remove the `curl | bash` and `Invoke-Expression` installation patterns. 2. Prefer installation through a trusted operating-system package manager or signed vendor package, with an explicitly pinned version. 3. If direct download is unavoidable: - Use an immutable, version-specific release URL. - Download the artifact to a local file without executing it. - Verify a separately published SHA-256 checksum and a cryptographic signature from a trusted signing key. - Present the verified file for inspection before executing it as a distinct step. 4. Keep CLI installation outside routine Skill execution and require explicit user approval before installing software. 5. Run the installer with the least-privileged account necessary and avoid administrator or root execution unless a documented installation step specifically requires it. 6. Document the files, directories, and permissions modified by the installer so users can assess its expected scope.
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
95% confidence
Finding
The skill description is intentionally broad: it instructs the agent to use this skill for ANY OOMOL Fusion API request, including reading, creating, updating, and deleting data. That broad trigger surface can cause the skill to be selected for many user prompts without narrowing conditions, increasing the chance that powerful write or destructive actions are invoked in contexts where safer scoping, extra confirmation, or more specific skill routing would be preferable.

Static analysis

No suspicious patterns detected.