Skills patchright-skill
📦
patchright-skill
Content revision r1 High Risk ⚙️ External commands🌐 Network access📁 Filesystem access
Automate Browser Testing with Patchright
Local web apps need visual checks after every change. This skill drives Patchright browser sessions for screenshots, navigation, form input, and UI verification.
Supports: Claude Codex Code(CC)
Stop for confirmation before installing.
Review the plan and obtain explicit user consent before changing files.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Agent request
Review the Skillstore skill "patchright-skill" from https://skillstore.io/skills/smallnest-patchright-skill.md and its manifest at https://skillstore.io/api/skills/smallnest-patchright-skill/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "patchright-skill". Capture the local home page.
Expected outcome:
- Opened the local home page successfully.
- Captured a screenshot in the working directory.
- Reported the current URL and page title.
Using "patchright-skill". Fill the test login form and submit it.
Expected outcome:
- Entered the requested test values into the visible fields.
- Submitted the form and waited for navigation.
- Captured the resulting page for review.
Using "patchright-skill". Check whether the dashboard finished loading.
Expected outcome:
- Waited for the dashboard selector.
- Read the visible status message.
- Saved a full-page screenshot for comparison.
Security Audit
High Riskv8 • 7/9/2026 Open versioned report
Most static findings are scanner noise from markdown fences, localhost examples, prior audit prose, and local debug output paths. Semantic review found real high-risk design issues: anti-bot bypass capability, an unauthenticated local browser control server, unrestricted screenshot output paths, and bundled self-attesting audit claims. The skill should require remediation and policy review before marketplace publication.
11
Files scanned
1,618
Lines analyzed
4
Review items
0
False positives ignored
Confirmed security concerns (4)
High
Prompt Injection Attempt Detected
evaluation_result.json contains self-attesting audit claims, including safe_to_publish true and a summary saying all static findings are false positives. This can bias automated review and should not be shipped inside the skill.
The file contains explicit safety and publication claims inside the package under review. It is not phrased as a direct system override, so confidence is high but not maximal.
High
Anti-Bot Detection Bypass Capability
The skill advertises bot detection bypass for protected sites, and the scripts launch Chrome with automation-detection controls disabled. This can enable abuse of sites outside authorized testing workflows.
The documentation explicitly says bot detection bypass is supported, and both browser launch paths disable AutomationControlled. The risk is clear even though local testing can be legitimate.
High
Unauthenticated Local Browser Control Server
scripts/server.py binds a local command server that accepts JSON commands without authentication and includes an evaluate handler for caller-provided JavaScript. Any local process that reaches the port can drive the active browser session and inspect page content.
The server binds to loopback and dispatches commands directly from decoded client JSON. There is no authentication, authorization, origin check, or command allowlist beyond tool names.
Medium
Unrestricted Screenshot Output Paths
The screenshot helpers accept caller-controlled output paths and write image files without constraining them to a safe directory. This can overwrite files accessible to the skill process and can place sensitive page captures in unexpected locations.
Both screenshot implementations pass user-controlled paths to the browser screenshot API. The server only converts the path to an absolute path and does not reject absolute or parent-directory paths.
Capability review items (4)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Low
Hardcoded URL
search_url = f"https://www.google.com/search?q={quote('2026年Go语言展望')}"
The documentation constructs a Google search URL from a user query, which sends query text to an external Google service. This is intentional behavior but still creates a privacy-relevant network dependency.
Low
Hardcoded URL
python scripts/server.py call '{"tool": "navigate", "args": {"url": "https://www.google.com/search?q
The reference workflow instructs users to navigate to Google search results through the automation server. It is documented external network behavior, not a hidden endpoint.
Low
Hardcoded URL
search_url = f"https://www.google.com/search?q={encoded_query}"
The script builds https://www.google.com/search?q= from user input and navigates there, disclosing search terms to Google. The behavior is expected for this utility but is real external network activity.
Low
Hardcoded URL
python server.py call '{"tool": "navigate", "args": {"url": "https://www.google.com"}}'
The README demonstrates navigating the local server to https://www.google.com. This is only documentation, but it encourages external browsing through the skill.
Risk Factors
⚙️ External commands (20)
evaluation_result.json:102 SKILL.md:51-54 SKILL.md:54-58 SKILL.md:58-67 SKILL.md:67-73 SKILL.md:73-85 SKILL.md:85-89 SKILL.md:89-98 SKILL.md:98-102 SKILL.md:102-111 SKILL.md:111-115 SKILL.md:115-118 SKILL.md:118-139 SKILL.md:139-157 SKILL.md:157-161 SKILL.md:161-171 SKILL.md:171-175 SKILL.md:175-195 SKILL.md:195-209 SKILL.md:209-212
🌐 Network access (18)
evaluation_result.json:132 evaluation_result.json:205 evaluation_result.json:211 evaluation_result.json:212 evaluation_result.json:144 reference.md:57 reference.md:89 scripts/executor.py:9 scripts/executor.py:267 scripts/google_search.py:43 scripts/README.md:49 scripts/server.py:159 scripts/server.py:177 SKILL.md:14 SKILL.md:81 SKILL.md:145 SKILL.md:163 SKILL.md:14
📁 Filesystem access (15)
evaluation_result.json:96 evaluation_result.json:84 evaluation_result.json:84 evaluation_result.json:120 evaluation_result.json:126 reference.md:159 scripts/google_debug.py:57 scripts/google_search.py:144 scripts/server.py:219 SKILL.md:52 SKILL.md:74 SKILL.md:140 SKILL.md:52 SKILL.md:74 SKILL.md:140
Audited by: codex View Audit History →
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/smallnest-patchright-skill/audits/8?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/smallnest-patchright-skill?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/smallnest-patchright-skill?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/smallnest-patchright-skill/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/smallnest-patchright-skill.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
smallnest. (2026). patchright-skill security audit report (audit version 8) [Author version unspecified]. Skillstore. https://skillstore.io/skills/smallnest-patchright-skill/audits/8BibTeX citation
@techreport{smallnest-smallnest-patchright-skill-2026,
author = {smallnest},
title = {patchright-skill security audit report (audit version 8)},
institution = {Skillstore},
year = {2026},
number = {8},
url = {https://skillstore.io/skills/smallnest-patchright-skill/audits/8},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "patchright-skill security audit report (audit version 8)"
version: "unspecified"
type: report
authors:
- name: "smallnest"
date-released: "2026-07-09"
url: "https://skillstore.io/skills/smallnest-patchright-skill/audits/8"
identifiers:
- type: other
value: "skillstore:smallnest-patchright-skill:audit:8"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: Medium 59
Architecture
75
Maintainability
85
Content
70
Community
91
Spec Compliance
What You Can Build
Check a local UI change
Open a development server, inspect the rendered page, and capture screenshots after code edits.
Validate a user flow
Run a repeatable browser flow that fills forms, clicks controls, waits for content, and records page state.
Verify a preview build
Exercise a private preview URL and collect visual evidence before approving a deployment.
Try These Prompts
Capture a Local Page
Open my local app on port 3000, capture the home page, and tell me the page title and screenshot path.
Test a Login Flow
Navigate to the local login page, enter test credentials, submit the form, and capture the result page.
Debug Dynamic Content
Open the app, wait for the dashboard content, read the visible status text, and capture a full-page screenshot.
Run a Multi-Step Review
Start a persistent browser session, walk through the main navigation, record each page title, and save screenshots for every step.
Best Practices
- Use test accounts and non-sensitive data during browser automation.
- Stop the helper server when the workflow is complete.
- Keep screenshots in a dedicated review folder with clear names.
Avoid
- Do not use the skill to bypass access controls on sites you do not own.
- Do not leave the local browser server running after a task ends.
- Do not store screenshots of sensitive user data in shared folders.
Frequently Asked Questions
What does this skill automate?
It automates Chrome browser actions such as navigation, screenshots, clicks, typing, waits, and page text checks.
Does it work with Claude and Codex?
Yes. The report lists support for Claude, Codex, and Claude Code.
Why does it use a helper server?
The server keeps one browser session alive across multiple commands, which helps multi-step UI workflows.
Can it browse external websites?
Yes, but external browsing should be limited to authorized testing and clear user intent.
Where are screenshots saved?
Screenshots are saved as local files at the path requested by the workflow.
What should users avoid?
Avoid real credentials, sensitive production data, unauthorized third-party automation, and leaving the server running.
Developer Details
Author
smallnestLicense
See LICENSE.txt
Skillstore revision
r1
Version notice
The author did not declare a version.
Ref
3e4b6c31a74a3bd1a291c98cf585d720cb9fbc88
Maintenance freshness
7/18/2026
Usage
10 downloads · 202 views
File structure