Back to skill

Security audit

Pixar Portrait Generator

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to be a simple external image-generation CLI, but users should review it because its advertised portrait-transform behavior, install path, Bash/network scope, and token handling are not tightly aligned or scoped.

Install only if you are comfortable sending prompts, reference identifiers, and a Neta API token to the documented external image API. Prefer a pinned or verified install source, avoid putting real tokens directly on the command line when possible, and treat this as text-prompt image generation rather than a confirmed selfie-to-portrait transformer.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
pixarportraitgenerator.js:14
Finding

API Token Exposure Through Command-Line Arguments

Content
View full analysis

Vulnerability Details

File Location: pixarportraitgenerator.js:14-37
Additional Locations: SKILL.md:17-21, 27-29; README.md:18-32, 55-67
Vulnerability Type: Sensitive credential exposure through process arguments and shell history
Risk Level: Medium

Vulnerable Code

javascript
function parseArgs(argv) {
  const args = { positional: [], size: "square", token: null, ref: null };
  for (let i = 0; i < argv.length; i++) {
    const a = argv[i];
    if (a === "--size") args.size = argv[++i];
    else if (a === "--token") args.token = argv[++i];
    else if (a === "--ref") args.ref = argv[++i];
    else args.positional.push(a);
  }
  return args;
}

async function main() {
  const argv = process.argv.slice(2);
  const { positional, size, token: tokenFlag, ref } = parseArgs(argv);

  const PROMPT = positional[0] || DEFAULT_PROMPT;
  const TOKEN = tokenFlag;

  if (!TOKEN) {
    console.error('\n✗ Token required. Pass via: --token YOUR_TOKEN');
    console.error('  Get yours at: https://www.neta.art/open/');
    process.exit(1);
  }

The documented invocation reinforces this practice:

bash
node pixarportraitgenerator.js "your description here" --token YOUR_TOKEN

Technical Analysis

The application exclusively accepts the Neta API token through the --token command-line option. Command-line arguments are not an appropriate secret transport mechanism because they can be exposed through:

  • Shell history files.
  • Process inspection tools while the process is running.
  • Process accounting and endpoint monitoring systems.
  • CI/CD job output or command tracing.
  • Wrapper scripts and automation logs.

The token is subsequently transmitted over HTTPS to the declared Neta API endpoint. The remote transmission itself is consistent with the documented image-generation functionality; the weakness is the local mechanism used to provide the credentia ...[truncated 1082 chars]

Remediation
View remediation

Remediation Suggestions

  1. Support a protected environment variable, such as NETA_TOKEN, instead of requiring the token on the command line:

    javascript
    const TOKEN = process.env.NETA_TOKEN;
    
  2. For interactive use, optionally read the token from a hidden stdin prompt so it is not echoed or recorded in shell history.

  3. If backward compatibility requires retaining --token, mark it as deprecated and emit a warning explaining the exposure risk.

  4. Update all documentation to recommend protected environment-variable or stdin-based input:

    bash
    export NETA_TOKEN="..."
    node pixarportraitgenerator.js "your description"
    
  5. Ensure error messages, diagnostic output, and exception handlers never include request headers or the token.

  6. Advise users who previously supplied tokens on the command line to clear affected histories and logs and rotate potentially exposed credentials.

T08 · Insecure Dependencies

Warning
Location
README.md:9
Finding

Unpinned Third-Party Installer and Skill Source

Content
View full analysis

Vulnerability Details

File Location: README.md:9-13
Additional Location: SKILL.md:32-34
Vulnerability Type: Mutable and unpinned supply-chain installation
Risk Level: Medium

Vulnerable Code

bash
npx skills add omactiengartelle/pixar-portrait-generator

The README also provides an installation command without an immutable version:

bash
clawhub install pixar-portrait-generator

Technical Analysis

The documented npx command can resolve and execute the current registry version of the skills package rather than a specifically reviewed version. The referenced skill source is likewise identified by a mutable repository-style name rather than an immutable commit, release digest, or verified checksum.

Consequently, the code installed or executed in the future may differ from the artifact covered by this audit. This does not prove that the current dependency or upstream source is malicious; it creates a supply-chain trust weakness in which compromise or unauthorized modification of the registry package, publisher account, repository, or distribution channel could affect later installations.

Attack Path

  1. An attacker compromises the package publisher, upstream source account, registry entry, or distribution channel.
  2. The attacker publishes or substitutes a modified installer or skill revision under the same mutable name.
  3. A user executes the documented npx skills add ... command or installs the latest ClawHub revision.
  4. The package manager resolves the attacker-controlled version because no reviewed version, commit, or integrity digest is pinned.
  5. The altered installer or installed skill executes with the privileges of the invoking user.

Impact Assessment

If the supply chain is compromised, malicious code could run with the permissions of the user performing installation. Depending on that user's privileges and the behavior of the substituted payload ...[truncated 370 chars]

Remediation
View remediation

Remediation Suggestions

  1. Pin the skills installer to a reviewed, exact version rather than allowing npx to resolve the latest release.
  2. Pin the skill source to an immutable commit hash, signed release, or content digest.
  3. Publish and verify cryptographic checksums or signed provenance for distributed artifacts.
  4. Use an official, verified package namespace and document the expected publisher identity.
  5. Prefer lockfile-backed or integrity-checked installation procedures where supported.
  6. Add release automation that verifies the packaged files correspond exactly to the reviewed repository commit.
  7. Document how users can inspect and verify the package before executing installation scripts.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (10)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The documented behavior claims a Pixar-style portrait generator using the Neta API and implies selfie/portrait transformation, but the analyzed behavior reportedly uses different API endpoints and functions more like a generic prompt-based generator. This mismatch is dangerous because users may disclose images, prompts, or tokens under false assumptions about what service is being contacted and what processing is occurring.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The manifest describes converting existing selfies, portraits, pets, and characters into Pixar-style renderings, implying image-to-image portrait transformation as the core purpose. The README instead states the skill generates portraits from text descriptions, with only an optional reference image for style inheritance, which is a materially different behavior.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding

The README instructs users to install and execute a package via npx skills without pinning a specific version. This creates a supply-chain risk: users may fetch and run whatever package version is current at execution time, including a compromised or malicious update. In a skill-install context, this is more dangerous because installation commands are likely to be copied and executed directly by users.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The usage examples consistently invoke the tool with natural-language prompts, and the --ref option is described only as a style-inheritance UUID rather than an input portrait to transform. This actively conflicts with the manifest's claim that the skill turns selfies and portraits into Pixar-style characters.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill declares tools: Bash and instructs users to invoke a Node script that depends on an external API token, but it does not constrain tool/network scope with explicit permissions or allowed-tools. In an agent environment, this increases the chance the skill can make unintended outbound requests or use broader execution capabilities than users expect, especially since the skill is marketed as a simple image generator.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The activation guidance says to use the skill whenever someone asks to generate Pixar-style portrait images, which is broad and lacks exclusions or tighter routing criteria. In agent systems, underspecified triggers can cause over-activation of a networked Bash-based skill in contexts where it is unnecessary, increasing accidental data exposure and unintended command execution.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The install instruction uses npx skills add omactiengartelle/pixar-portrait-generator without pinning a version or immutable reference. This creates a supply-chain risk because future changes to the package or resolver target could cause users to install unexpected or malicious code under the same name.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

The hardcoded external endpoint confirms that the skill is designed to transmit content to a third-party domain. Even though this is expected for an image-generation integration, it creates privacy and data-handling risk because prompts and related identifiers leave the trusted boundary and are processed by an external provider.

Content

Scanner excerpt · pixarportraitgenerator.js (reported line 62)May include surrounding context.

js
console.error(`→ Submitting task (${dims.width}×${dims.height})...`);

  const submitRes = await fetch("https://api.talesofai.com/v3/make_image", {
    method: "POST",
    headers,
    body: JSON.stringify(body),

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

The hardcoded external endpoint confirms that the skill is designed to transmit content to a third-party domain. Even though this is expected for an image-generation integration, it creates privacy and data-handling risk because prompts and related identifiers leave the trusted boundary and are processed by an external provider.

Content

Scanner excerpt · pixarportraitgenerator.js (reported line 62)May include surrounding context.

js
console.error(`→ Submitting task (${dims.width}×${dims.height})...`);

  const submitRes = await fetch("https://api.talesofai.com/v3/make_image", {
    method: "POST",
    headers,
    body: JSON.stringify(body),

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The polling request repeatedly contacts the same external service and includes the authentication header, extending the period during which secrets and metadata are exposed in transit and logs. While this call mainly retrieves task status rather than new user content, it is still a genuine external communication channel to a third party.

Content

Scanner excerpt · pixarportraitgenerator.js (reported line 94)May include surrounding context.

js
await new Promise((r) => setTimeout(r, 2000));

    const pollRes = await fetch(
      `https://api.talesofai.com/v1/artifact/task/${taskUuid}`,
      { headers }
    );