Back to skill

Security audit

Watercolor Art Generator

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward image-generation skill that uses a disclosed external API, with privacy and token-handling caveats but no hidden persistence or destructive behavior.

Before installing, prefer the ClawHub install path or a pinned, verified installer instead of the unpinned npx command. Treat prompts, reference image UUIDs, and the Neta token as data sent to the remote API, and avoid using confidential content. Pass tokens only in environments where command-line history and process listings are not exposed, or rotate the token afterward if it may have been exposed.

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)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:32
Finding

Unpinned Third-Party Installation Command Creates a Supply-Chain Risk

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:32-34; README.md:12-15
Vulnerability Type: Unpinned third-party package and skill installation
Risk Level: Medium

Vulnerable Code

SKILL.md:32-34:

bash
## Install
```bash
npx skills add blammectrappora/watercolor-art-generator
text

`README.md:12-15`:

```bash
**Via npx skills:**
```bash
npx skills add blammectrappora/watercolor-art-generator
text

### Technical Analysis

The documented installation command invokes `npx` without pinning an exact version of the `skills` package. It also identifies the skill by a mutable owner/name reference rather than a fixed release, commit, or content-integrity hash.

`npx` may download and execute package code from the configured npm registry. Consequently, the code executed during installation can differ from the artifact covered by this audit. Compromise of the npm package, registry resolution, publisher account, or remote skill release could introduce attacker-controlled installation behavior.

This is a supply-chain exposure rather than evidence that the currently audited files contain malicious code.

### Attack Path

1. An attacker compromises the publisher account, package, registry resolution path, or mutable remote skill release.
2. The attacker publishes a modified package or skill containing malicious installation logic.
3. A user follows the documented `npx skills add` command.
4. `npx` retrieves the current mutable package version rather than a reviewed and integrity-pinned version.
5. The malicious package or installation logic executes with the privileges of the invoking user.

### Impact Assessment

Successful exploitation could allow arbitrary code execution under the account running the installation command. Depending on that account's permissions, the attacker could access user files and credentials, modify installed skills, initiate network requests, or establish fur
...[truncated 129 chars]
Remediation
View remediation

Remediation Suggestions

  • Pin the skills CLI to an exact reviewed version, for example by specifying an immutable package version.
  • Pin the installed skill to an exact release or commit rather than a mutable owner/name reference.
  • Use registry integrity hashes and a lockfile where supported.
  • Document the expected package registry and publisher identity.
  • Prefer installation from a reviewed local artifact when the audited files are already available.
  • In controlled environments, preinstall the verified CLI instead of allowing npx to download and execute packages dynamically.
  • Verify package signatures or checksums before installation where the ecosystem supports them.

T09 · Insecure Skill Coding Practices

Warning
Location
watercolorartgenerator.js:3
Finding

API Token Exposed Through Process Command-Line Arguments

Content
View full analysis

Vulnerability Details

File Location: watercolorartgenerator.js:3-18, 25-26, 43-47; SKILL.md:15-19, 25-27; README.md:23-26, 61-81
Vulnerability Type: Sensitive credential passed through process arguments
Risk Level: Medium

Vulnerable Code

watercolorartgenerator.js:3-18:

javascript
const args = process.argv.slice(2);

let prompt = null;
let size = 'portrait';
let tokenFlag = null;
let refUuid = null;

for (let i = 0; i < args.length; i++) {
  if (args[i] === '--size' && args[i + 1]) {
    size = args[++i];
  } else if (args[i] === '--token' && args[i + 1]) {
    tokenFlag = args[++i];
  } else if (args[i] === '--ref' && args[i + 1]) {
    refUuid = args[++i];
  } else if (!args[i].startsWith('--') && prompt === null) {
    prompt = args[i];
  }
}

watercolorartgenerator.js:25-26:

javascript
const TOKEN = tokenFlag;

watercolorartgenerator.js:43-47:

javascript
const HEADERS = {
  'x-token': TOKEN,
  'x-platform': 'nieta-app/web',
  'content-type': 'application/json',
};

README.md:61-81 documents the same credential-handling pattern:

bash
node watercolorartgenerator.js "your prompt" --token YOUR_TOKEN
bash
export NETA_TOKEN="your_token_here"
node watercolorartgenerator.js "your prompt" --token "$NETA_TOKEN"

Technical Analysis

The application requires the API token to be supplied using the --token command-line option and reads it directly from process.argv. Command-line arguments are not an appropriate secret transport mechanism because they may be recorded in shell history and exposed through process inspection, diagnostics, monitoring agents, crash reports, or command telemetry.

The documented environment-variable example does not remove this exposure. Expanding "$NETA_TOKEN" into the command still places the resulting secret value in the child process' ...[truncated 1415 chars]

Remediation
View remediation

Remediation Suggestions

  • Read the token directly from process.env.NETA_TOKEN without expanding it into the command line:
    javascript
    const TOKEN = process.env.NETA_TOKEN;
    
  • Update examples to use:
    bash
    NETA_TOKEN="your_token_here" node watercolorartgenerator.js "your prompt"
    
  • For interactive use, support reading the token from a hidden terminal prompt or standard input.
  • For automated environments, integrate with a permission-restricted secret manager or credential file.
  • Deprecate and eventually remove the --token option.
  • If temporary backward compatibility is necessary, display a warning when --token is used.
  • Ensure tokens are redacted from logs, errors, telemetry, diagnostics, and crash reports.
  • Document token rotation and revocation procedures in case command-line exposure has already occurred.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README describes generating art through an external API but does not clearly warn users that their prompts and optional reference-image identifiers are transmitted to a third-party service. This can lead users to unknowingly disclose sensitive personal, proprietary, or confidential information in prompts or linked assets, especially because the skill markets convenience and accepts free-form user input.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

The install instruction uses npx skills add blammectrappora/watercolor-art-generator without pinning a specific package or tool version. This can cause users to execute whatever version of the skills package resolves at install time, creating a supply-chain risk if a malicious or compromised release is published later; because this skill explicitly tells users to run the command, the documentation meaningfully increases the chance of unsafe execution.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
97% confidence
Finding

This request transmits user-supplied prompts and the provided x-token to an external domain, api.talesofai.com. In isolation external API use can be expected for an image-generation skill, but in this context it is dangerous because the destination contradicts the declared provider, so sensitive data and credentials may be exfiltrated to an undisclosed service.

Content

Scanner excerpt · watercolorartgenerator.js (reported line 106)May include surrounding context.

js
};
  }

  const makeRes = await request('POST', 'https://api.talesofai.com/v3/make_image', makeBody);

  const taskUuid = typeof makeRes === 'string' ? makeRes : makeRes.task_uuid;
  if (!taskUuid) {

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill metadata explicitly claims it uses the Neta AI API, but the implementation sends the user's prompt and bearer-like token to api.talesofai.com instead. This is a serious trust-boundary violation because users may disclose sensitive prompts or credentials under false assumptions about the receiving service, and the mismatch strongly suggests deception or at minimum unsafe misrepresentation.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The polling call continues communicating with the same undisclosed external service, leaking the generated task identifier and confirming ongoing interaction with that provider. While the task UUID alone is less sensitive than the initial prompt and token, it reinforces the unauthorized third-party data flow and expands the exposure window.

Content

Scanner excerpt · watercolorartgenerator.js (reported line 114)May include surrounding context.

js
process.exit(1);
  }

  const pollUrl = `https://api.talesofai.com/v1/artifact/task/${taskUuid}`;
  const maxAttempts = 90;

  for (let i = 0; i < maxAttempts; i++) {