Back to skill

Security audit

ClassMarker

Security checks for vulnerabilities and agentic risk

Overview

The ClassMarker functions fit the stated purpose, but setup directs the agent to run downloaded installation code without a clear approval step.

Review the setup before enabling it. Install a reviewed CLI version yourself, approve any changes to other installed agent skills, and use a ClassMarker key limited to the data needed. OOMOL is the disclosed intermediary for credentials and requests. The downloaded installer and hosted connector implementation were not verified; no malicious behavior was established in the bundled skill.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (5)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied code performs remote installer execution rather than searching or reading ClassMarker data. Although installation can be a supporting implementation detail, the supplied content establishes no relationship between this installer and the declared ClassMarker functionality. The downloaded script's further behavior is not visible.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The supplied code performs only a login operation, rather than the declared ClassMarker search and reading functions. Authentication could be a supporting implementation detail, but the chunk provides no evidence that this login supports ClassMarker access.

Content

No source excerpt is available for this finding.

External Script Fetching

High
Category
Supply Chain
Confidence
99% confidence
Finding

The curl-to-bash command executes the installer endpoint's current response with the agent's local user privileges, without a pinned version or independently verified integrity. Compromise of the distribution endpoint could therefore lead to arbitrary code execution and access to local files or credentials; HTTPS does not protect against a compromised publisher, and the PowerShell alternative has the same underlying risk.

Content

Scanner excerpt · SKILL.md (reported line 59)May include surrounding context.

The setup workflow executes downloaded installer scripts without warning the user about remote code execution or system changes.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

A missing CLI triggers instructions to install software by immediately executing downloaded scripts, without requiring informed user approval for local code execution and system changes. Although setup is conditional and relevant to the connector, a read-oriented ClassMarker request does not itself authorize installing software on the user's machine.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 80)May include surrounding context.

md
- **`OOMOL_INSUFFICIENT_CREDIT`** — OOMOL billing stop. Recharge at `https://console.oomol.com/billing/token-recharge` before retrying.

- **HTTP 402 without `OOMOL_INSUFFICIENT_CREDIT`** — the third-party account itself has no active plan or credit. Report the provider error message as is; do not ask the user to recharge OOMOL.

## Resources