Back to skill

Security audit

Databox

Security checks for vulnerabilities and agentic risk

Overview

Databox changes require your approval, but setup can execute a downloaded installer without a separate approval or integrity check.

Review and approve CLI installation separately, including any bundled agent-skill changes. Prefer a verified release and limit the connected Databox account's permissions. Keep the skill's approval requirements for writes and deletions. The artifact review resolved the scanner's missing-reference gaps, but could not establish the downloaded installers' behavior; no malicious intent was demonstrated.

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 (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied command runs an Oomol-hosted installer rather than performing Databox operations. Nothing in the supplied chunk establishes that this installation supports the declared Databox purpose. The downloaded script's contents are not provided, so its further behavior cannot be assessed.

Content

No source excerpt is available for this finding.

External Script Fetching

High
Category
Supply Chain
Confidence
98% confidence
Finding

The command pipes a mutable remote installer directly into bash, allowing its contents to execute with the agent's local permissions without prior verification. HTTPS protects transport but does not prevent a compromised publisher or distribution endpoint from serving malicious code that could access local files and credentials; this is the same underlying installation risk identified by SQP-2.

Content

Scanner excerpt · SKILL.md (reported line 61)May include surrounding context.

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

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The missing-CLI fallback directs the agent to execute downloaded shell or PowerShell code without explicit installation approval or integrity verification. A Databox request could therefore trigger local software installation and arbitrary code execution under the agent's account if the installer or its delivery infrastructure is compromised; the missing-CLI condition limits exposure but does not eliminate it.

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 82)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