Back to skill

Security audit

id-photo-generator

Security checks for vulnerabilities and agentic risk

Overview

ClawScan’s verdict is benign with medium confidence: the scanners found no AIG issues, while Skillspector reported inconsistencies in identity-photo authenticity and image-size checks that the verdict treats as purpose-related concerns.

The verdict is benign, but consider the reported validation gaps before installing, especially if consistent authenticity checks or minimum image dimensions matter to your use. The overall confidence is medium; the scanners report complete component coverage, though Skillspector’s output includes an exit-status error and serializer warnings.

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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Ordinary mode permits generation from suspected recaptured or synthetic photos despite the manifest's unconditional prohibition.

High
Category
Excessive Agency
Confidence
99% confidence
Finding

The manifest unconditionally prohibits generating photos from suspected recaptured or synthetic inputs, but ordinary mode explicitly allows them. This creates a policy downgrade path in an identity-photo workflow; preview warnings reduce user confusion but do not prevent reuse of the resulting image.

Content

No source excerpt is available for this finding.

Normal-mode test expectations allow rephotographed ID photos despite the manifest's unconditional rejection requirement.

High
Category
Excessive Agency
Confidence
99% confidence
Finding

The test checklist explicitly expects ordinary-mode generation to proceed for rephotographed images, although the manifest prohibits it. These expectations can institutionalize the bypass by treating a fail-open implementation as correct and a stricter implementation as a regression.

Content

No source excerpt is available for this finding.

Ordinary mode permits suspected rephotographed or composited portraits despite the manifest's unconditional rejection requirement.

High
Category
Excessive Agency
Confidence
99% confidence
Finding

The validation table explicitly allows suspected recaptured, composite, and deepfake inputs through ordinary mode with only a warning. This directly weakens the manifest's authenticity gate in an identity-photo generator, although the skill is not itself demonstrated to be an identity-authentication system.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The ordinary-mode messaging policy permits generation for suspected rephotographed or composite images, contradicting the manifest's unconditional rejection requirement.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The messaging policy categorizes suspected recaptured or composite images as eligible for successful generation with a warning, conflicting with the manifest's mandatory rejection. Because this reference guides agent behavior, it can reinforce a fail-open authenticity decision; the warning does not technically restrict subsequent use.

Content

No source excerpt is available for this finding.

The documented hard resolution gate is implemented as a soft resize that can also bypass the minimum dimension requirement.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

Automatic resizing above 4096 pixels is explicitly documented, so that part of the finding is not a policy violation. However, the minimum dimension is checked only before resizing: an input such as 512×8192 becomes 256×4096 and can pass the remaining checks despite violating the minimum-resolution requirement.

Content

No source excerpt is available for this finding.