Back to skill

Security audit

GenArt

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local generative-art tool with some under-documented SVG handling and resource-risk edges, but no evidence of deception, persistence, exfiltration, or privilege escalation.

Install only if you are comfortable running a local Python art generator that writes files and may call the system SVG renderer for PNGs. Avoid feeding it untrusted SVGs, avoid extreme size/count/layer arguments, and choose output paths carefully because it can overwrite files at paths the current user can write.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
genart.py:660
Finding

Untrusted SVG Content Is Automatically Passed to a Native Renderer

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
genart.py:384
Finding

Unbounded Generation Parameters Permit Resource Exhaustion

Content
View full analysis
= count: break ``` ```python def gen_waves(seed, width, height, layers, palette_name): rng = random.Random(seed) palette, pname = get_palette(palette_name, rng) # Background sky gradient defs = svg_linear_gradient("sky", 0, 0, 0, 100, [ (0, palette[0], 1), (50, palette[1], 1), (100, ...[truncated 4852 chars]
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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
Findings (5)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The documented purpose does not fully disclose behavior: the skill reportedly reads and modifies input SVGs for glitch art, exposes undeclared modes, and invokes an external binary for rendering. This mismatch is dangerous because hidden file-processing and shell-adjacent behavior can surprise reviewers and users, increasing the risk of unsafe file access, command execution pathways, or abuse through crafted inputs.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill advertises code-capable behavior including file read, file write, and shell execution, but does not declare any explicit tool scope or permissions. This weakens containment and reviewability, making it easier for the skill to access files or invoke commands beyond what a user would reasonably expect from the manifest alone.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The glitch mode reads an arbitrary local file path supplied by the user and parses its contents as SVG, which expands the skill from art generation into local file access and transformation. In an agent context, this can be abused to read sensitive local files or probe file existence, and if the selected file is later written back or converted, it may also propagate untrusted content through downstream processors.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The manifest says the skill generates algorithmic visual art with SVG and PNG output, but does not disclose that PNG output depends on spawning an external executable. Executing a host binary is a stronger capability than plain image generation and is context-sensitive because it reaches beyond in-process rendering into system-level command execution.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · genart.py (reported line 734)May include surrounding context.

python
rsvg = "/usr/bin/rsvg-convert"
    if os.path.exists(rsvg):
        try:
            subprocess.run([rsvg, svg_path_out, "-o", png_path],
                           check=True, capture_output=True, timeout=30)
            files.append(png_path)
        except (subprocess.CalledProcessError, subprocess.TimeoutExpired, OSError):