Security audit
GPT Image 2.6 AI 图像生成 | 文生图 图生图 智能修图 创意图片制作
Security checks for vulnerabilities and agentic risk
Overview
The skill describes a proportionate image-generation workflow that requires user confirmation before spending credits.
Expect prompts and selected reference images to be sent to Qinghu, with generation charges requiring confirmation. Verify the CLI package and any QHKIT_BASE_HOST override before providing a token. The external CLI implementation was not audited; the scanner's missing JSON references were verified as files created during normal use.
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
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- 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