Back to skill

Security audit

Process-Diagram-Pro

Security checks for vulnerabilities and agentic risk

Overview

The diagram generator is mostly local and purpose-aligned, but its instructions add broad browser automation and an unpinned global npm install that are outside the stated diagram purpose and under-scoped.

Review this skill before installing. Use it in an isolated environment, pin Python dependencies, and avoid running the documented global agent-browser install unless you specifically need browser automation and can contain it. Do not use the browser commands on sensitive pages or accounts without explicit control over screenshots, extracted text, and form actions.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:162
Finding

Unpinned Global Installation of an Unrelated npm Package

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 162-163
Vulnerability Type: Unpinned third-party dependency installed globally
Risk Level: Medium

bash
npm install -g agent-browser
agent-browser install --with-deps

Technical Analysis

The skill documentation instructs the agent to install the latest available version of agent-browser globally and then install its additional dependencies. The package is not declared in the project dependency manifest, is not pinned to a reviewed version or integrity hash, and is unrelated to the core local diagram-rendering functionality.

npm installations can execute package lifecycle scripts such as preinstall, install, and postinstall. Consequently, the behavior executed by this instruction can change without any modification to the audited project: a compromised package release, compromised maintainer account, malicious transitive dependency, or future unsafe release could execute arbitrary code during installation. Global installation also expands the affected scope by placing commands and package files in shared system or user-level npm locations.

The subsequent agent-browser install --with-deps command may retrieve and install further browser components or operating-system dependencies whose versions and integrity are not controlled by this repository.

No evidence was found that the currently named package is malicious. The vulnerability is the unsafe, unpinned, globally scoped dependency-installation procedure.

Attack Path

  1. An attacker compromises the agent-browser package, one of its transitive dependencies, its publishing account, or a future package release.
  2. The agent follows SKILL.md and runs npm install -g agent-browser.
  3. npm retrieves the attacker-controlled version because no exact version or integrity constraint is specified.
  4. Malicious package code executes through an npm lifecycle script or when the installed CL ...[truncated 1090 chars]
Remediation
View remediation

Remediation Suggestions

  1. Remove the browser-installation section if browser automation is not essential to generating diagrams.
  2. Do not direct agents to install packages globally. Use an isolated, least-privileged project environment instead.
  3. If the dependency is required, declare an exact reviewed version, for example through a project-local package.json and lockfile, rather than installing the latest release.
  4. Commit and verify a lockfile with integrity metadata. Use deterministic installation such as npm ci.
  5. Disable lifecycle scripts where operationally possible using npm ci --ignore-scripts, and separately review any required installation scripts before enabling them.
  6. Pin and verify browser binaries and other components retrieved by agent-browser install --with-deps.
  7. Run browser tooling in a sandbox or container without access to secrets, unrelated user files, privileged sockets, or administrative permissions.
  8. Document why the dependency is necessary and restrict its allowed network destinations to trusted documentation sites.
  9. Add automated dependency scanning and periodically review both direct and transitive dependencies for known vulnerabilities and ownership changes.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (27)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger description says users can directly describe needs and the AI will automatically invoke the skill, with examples phrased very broadly around drawing or generating diagrams. Overly broad activation criteria increase the chance the skill is invoked in unintended contexts, which is more concerning here because the same skill also documents unrelated browsing capabilities.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill is presented as a process diagram generator, but it also documents broad agent-browser capabilities for opening arbitrary websites, clicking elements, filling forms, taking screenshots, and extracting page text. This materially expands the skill’s effective scope into general web browsing and scraping, which can expose users to unintended data access, prompt injection from webpages, and exfiltration of sensitive page content.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documentation includes commands to open external pages, capture screenshots, extract text, and interact with page elements, but it does not warn about the risks of visiting untrusted sites, handling sensitive content, or leaking information through screenshots and copied text. In an agent context, this omission can lead to unsafe browsing behavior and accidental collection of confidential data.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The documentation creates a scope mismatch: it claims the skill is for chemical/process diagram generation, then later states it supports browser-based web scraping and research. This kind of hidden or secondary capability is dangerous because users and orchestration systems may trust the skill for a narrow purpose while it actually enables broader network interaction and content acquisition.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code file contains user-facing natural-language content in docstrings and usage examples that consistently force Chinese as the only language. Under the policy, language constraints should offer user opt-in or be clearly justified as region-specific; neither is present here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This code file contains natural-language docstrings and usage examples exclusively in Chinese, including the module description and API example text. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is clearly justified, which is not stated here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

This file contains natural-language descriptions, docstrings, and usage guidance exclusively in Chinese, including the module title and example usage. That can violate a language/locale policy when users are not given an opt-in choice or any explanation that the skill is intentionally region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This Python file contains natural-language documentation and runtime messages primarily in Chinese, including the module description and examples, with no indication that language selection is configurable or that the skill is intentionally region-specific. Under the policy rule, forcing a specific language without user opt-in is a locale/language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The file’s natural-language descriptions are entirely in Chinese, including the module header and usage documentation, with no indication that the skill supports other languages or that Chinese is a required locale. Under the policy, forcing a specific language without user opt-in or clear regional justification is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This code file contains natural-language documentation and comments stating the module is for automatically finding Chinese fonts and using Chinese-specific font priorities. Under the policy, forcing a specific language/locale without user opt-in is a violation unless the restriction is clearly justified as region-specific, which is not documented here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The heading emphasizes a Chinese-language specialization and presents the skill primarily in Chinese, which may imply a fixed language/locale experience. There is no explicit statement that users may choose another language or locale for outputs.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This code file contains natural-language documentation entirely in Chinese while not indicating that the skill is region-specific or that users may choose another language. Under the stated policy, forcing a specific language without opt-in is a locale-policy concern even when it appears in comments or docstrings.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

This code file contains user-facing natural-language documentation entirely in Chinese, including the class and initializer descriptions, without any indication that language choice is optional or that the skill is intentionally region-specific. Under the stated policy, forcing a specific language without opt-in can be a locale-policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

This code file consistently documents the skill only in Chinese, including class descriptions and user-facing parameter semantics, with no indication that language is configurable or optional. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code file's natural-language docstrings and comments are entirely in Chinese, including the module description and usage-oriented class documentation. Under the stated policy, forcing a specific language without user opt-in can be a locale-policy violation unless the constraint is explicitly documented and justified, which is not present here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

This code file includes natural-language strings and documentation primarily in Chinese, with some English labels, but there is no indication that the user can choose the language or that the locale restriction is intentional. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The natural-language content in the module and class documentation is entirely in Chinese, including user-facing descriptions and argument explanations. Under the policy rule, forcing a specific language without opt-in can be a locale-policy issue when no alternative or justification is provided.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This file's primary natural-language documentation and class/method descriptions are presented in Chinese, which can force a language expectation on users or maintainers who did not opt into that locale. The policy calls for flagging language/locale constraints when no choice or justification is provided.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The module docstring states a third-priority fallback of English fonts such as DejaVu Sans, implying those fonts are part of the actual selection chain. In code, ENGLISH_FONTS is never used, and get_font() falls back directly to ImageFont.load_default(), so the documented fallback behavior does not match the implementation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The diagram content includes the Chinese label "Data看板" alongside English labels, which introduces a locale-specific term without any indication that multilingual output is optional or user-selected. This can violate a language/locale policy when the skill implicitly chooses a language for generated artifacts.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The test section says it is performing a drawing test and reports success, but it calls ImageDraw.Draw at L231 without importing ImageDraw anywhere in the module. This contradicts the apparent intent of the self-test because the test will raise a NameError before saving the image.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The main test path performs a filesystem write by saving an image to /tmp/font_loader_test.png. Although this is test code, it is still a file write in a code file and there is no prior confirmation or explicit warning in the surrounding documentation that running the module will create an output file.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
95% confidence
Finding

The dependency is specified with a minimum version only, which allows future unreviewed releases to be installed. This weakens supply-chain reproducibility and can expose consumers to breaking or vulnerable versions introduced after the skill was authored.

Content

Scanner excerpt · requirements.txt (reported line 1)May include surrounding context.

text
matplotlib>=3.7.0
pillow>=10.0.0
numpy>=1.24.0

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
98% confidence
Finding

Pillow is unpinned, so installations may resolve to different versions over time, including versions with known security issues or future regressions. Because Pillow processes image data and has a history of parser-related vulnerabilities, leaving it unpinned increases supply-chain and exposure risk.

Content

Scanner excerpt · requirements.txt (reported line 2)May include surrounding context.

text
matplotlib>=3.7.0
pillow>=10.0.0
numpy>=1.24.0

Unverifiable Dependency: pillow has 16 known advisory(ies) (CVE-2016-2533 (Pillow buffer overflow in ImagingPcdDecode); CVE-2023-50447 (Arbitrary Code Execution in Pillow); CVE-2021-27922 (Pillow Uncontrolled Resource Consumption) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
90% confidence
Finding

The manifest does not pin Pillow, and Pillow has multiple known advisories across versions, so it is impossible to verify from this file whether deployments will install a safe release. In a skill that may process untrusted image files, this uncertainty is more dangerous because vulnerable image parsers can lead to denial of service or potentially code execution.

Content

No source excerpt is available for this finding.