Back to skill

Security audit

Octen

Security checks for vulnerabilities and agentic risk

Overview

This Octen integration needs review because its setup can execute changing remote installers without clear user approval.

Review and approve a pinned, verified CLI installation separately, including any additional agent skills it installs. Confirm intended generation inputs and credit use instead of relying on the read-only labels. The bundled instructions show no malicious intent, but downloaded installers, CLI runtime and live connector behavior are outside this artifact review.

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
97% confidence
Finding

The supplied code performs remote installer execution rather than Octen data operations. Installing a dependency could be a supporting implementation detail, but nothing supplied establishes that the Oomol installer supports Octen. The remote script's contents and any further effects cannot be determined from this chunk.

Content

No source excerpt is available for this finding.

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

Piping a live HTTPS response directly into bash executes remote code without a pinned version or independent integrity verification. Compromise of the distribution endpoint or its delivery pipeline could yield arbitrary execution with the agent's privileges; the endpoint's legitimate dependency role does not eliminate that supply-chain risk, and this overlaps with the setup-consent finding.

Content

Scanner excerpt · SKILL.md (reported line 72)May include surrounding context.

Image generation is classified as a read, bypassing the documented confirmation requirement.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The skill declares all untagged actions to be reads that are safe to run directly, but generate_images creates or edits images and returns generated artifacts. Leaving it untagged bypasses the skill's confirmation requirement for state-changing operations and can permit unintended generation or billable resource consumption; the user's overall request may authorize generation, but the documented classification is still incorrect.

Content

No source excerpt is available for this finding.

The fallback setup instructs the agent to download and immediately execute remote installation scripts without warning the user about the system changes or obtaining approval.

Medium
Category
Output Handling
Confidence
97% confidence
Finding

A missing CLI triggers instructions to install executable software without first obtaining approval for local system changes. Although this is a conditional dependency-setup path rather than routine behavior, executing the downloaded installer could modify files and configuration or access data with the agent process's privileges.

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 93)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