ClawHub
Back to skill

Security audit

Chatgpt Apps

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent ChatGPT app-building guide, with expected but important care needed around secrets and deployment.

Install only if you are comfortable with an agent helping create and deploy a ChatGPT app. Do not paste real client secrets, Supabase service role keys, or other production credentials into chat or committed files; use environment variables or a secrets manager, and review deployment targets and public URLs before publishing.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The database setup section directs use of a Supabase service role key without emphasizing that it is a highly privileged server-side secret capable of bypassing normal client restrictions. In a skill that generates app code and deployment steps, failing to warn about this can lead to accidental inclusion in client code, chat transcripts, logs, or repositories, resulting in full database compromise.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill includes a deploy command and production deployment workflow but does not warn users that running deployment can publish a live service, expose endpoints publicly, and make connected integrations reachable from the internet. In an agent-driven context, omission of this warning increases the chance a user triggers deployment without understanding external exposure or associated data/security consequences.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The authentication section instructs users to gather and configure credentials such as client IDs and client secrets, but it does not warn against sharing secrets in chat, logs, generated files, or version control. Because this skill is designed for conversational app building, users may paste sensitive values directly into the conversation or generated artifacts, creating a realistic path to credential leakage.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.