ClawHub
Back to skill

Security audit

MV Image Forensics

Security checks for vulnerabilities and agentic risk

Overview

This local image-comparison skill uses proportionate file access, but its comparison labels and conclusions about blur can be unreliable.

Account for the three Python dependencies and choose separate output paths. Check original images and actual scaling before trusting the comparison labels or conclusions about who caused blur; matching a reproduced resize does not establish that resizing preserved quality.

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
  • 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 (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The comparator can enlarge images while the documentation promises shrink-only normalization and its default output title says they were not enlarged. In a skill used to make forensic quality decisions, this discrepancy can cause users or agents to treat resampled evidence as an original-pixel comparison; the inaccurate dependency claim is a separate packaging defect, not itself a demonstrated security vulnerability.

Content

No source excerpt is available for this finding.

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

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 attribution verdict can blame the generator after normalizing away pipeline-induced resampling damage.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The script can resize the source to reproduce the delivery pipeline and then interpret a match against that transformed source as proof that postprocessing caused no damage. A delivery image softened by exactly that resize can therefore receive a generator-blaming verdict; additionally, the 95% equality threshold does not justify claiming zero damage to all character pixels.

Content

No source excerpt is available for this finding.

The comparator can enlarge source images despite documenting shrink-only normalization and labeling its output as not enlarged.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The scale factor is not capped, so a detected or manually supplied character height below the requested target causes enlargement. Using INTER_AREA does not preserve original pixels during enlargement, and the default 'not enlarged' title conceals that transformation, undermining the integrity of the resulting forensic comparison.

Content

No source excerpt is available for this finding.