Back to skill

Security audit

AI Co-Founder Engine (Scored Startup Builder)

Security checks for vulnerabilities and agentic risk

Overview

This startup-planning skill is coherent and purpose-aligned, with the main caveat that its market-research fallback can send search queries to DuckDuckGo.

Install only if you are comfortable with startup idea or market-research queries potentially being sent to external web search. Avoid including confidential business plans, secrets, customer data, or unreleased proprietary details in prompts that may trigger research.

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
  • 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
Category
Not specified by scanner
Confidence
92% confidence
Finding

The README describes the skill in very broad terms ('Just describe your idea') and lists generic startup-assistance capabilities without clear scope boundaries. This can cause the skill to trigger on a wide range of ordinary business or planning requests, creating ambiguous routing and increasing the chance that it intercepts requests better handled by other skills or applies startup-analysis behavior in unintended contexts.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

This code performs an outbound fetch to a third-party endpoint using user-controlled input as the query parameter. In an agent skill context, that is dangerous because tool inputs can include sensitive user data or internal context, and the fallback behavior may trigger unexpectedly when a preferred local/user tool is unavailable.

Content

Scanner excerpt · scripts/tool_router.js (reported line 19)May include surrounding context.

js
// fallback (no API key required, but basic)
async function fallbackSearch(query) {
  const res = await fetch(`https://api.duckduckgo.com/?q=${encodeURIComponent(query)}&format=json`);
  const data = await res.json();

  return {

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The fallbackSearch function sends the caller-provided query to DuckDuckGo over the network without any indication in this code that the user consented to external transmission. If input may contain sensitive prompts, secrets, internal data, or user-provided content, this creates a real privacy and data-leak risk because the fallback path silently exports data to a third party.

Content

No source excerpt is available for this finding.