ClawHub
Back to skill

Security audit

oracle-coordination

Security checks for vulnerabilities and agentic risk

Overview

This is a simple task-routing skill with disclosed example code; it has broad wording and an optional external API example, but no hidden execution, persistence, or destructive behavior.

Before installing, treat this as a lightweight routing template. Narrow when it should activate, and only use the API example with a trusted endpoint after redacting sensitive tickets, prompts, incidents, or customer data.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
96% confidence
Finding
The manifest description says the skill 'Routes any task to the right specialist agent automatically,' which is extremely broad and lacks boundaries on what kinds of requests should invoke it. Because it claims applicability to 'any task' without exclusions or a constrained context, it risks unintended invocation from common everyday requests.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The overview again states 'Routes any task to the right specialist agent automatically,' without defining activation conditions, trigger phrases, or non-applicable cases. This makes the invocation boundary unclear and increases the chance of accidental routing on general conversation.

External Transmission

Medium
Category
Data Exfiltration
Content
## x402 Integration (Pay-Per-Call)

```bash
curl -X POST https://your-api.com/api/oracle/route \
  -H "Content-Type: application/json" \
  -H "X-PAYMENT: <base64_payment>" \
  -d '{"task":"Fix the security vulnerability"}'
Confidence
87% confidence
Finding
The skill includes an example that sends task content to an external API endpoint, which can expose sensitive prompts, tickets, incident details, or internal security data to a third party. In this skill's context, routing may involve security, support, or operational requests, so transmitting raw task text externally increases the chance of data leakage and unintended disclosure.

Static analysis

No suspicious patterns detected.