Skills building-chatgpt-apps
📦
building-chatgpt-apps
Content revision r1 Medium Risk ⚙️ External commands🌐 Network access🔑 Env variables📁 Filesystem access
Build ChatGPT Apps with Widgets
Building ChatGPT Apps requires specific widget metadata, server structure, and testing steps. This skill provides templates and guidance for creating interactive Apps with MCP servers and embedded widgets.
Supports: Claude Codex Code(CC)
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Agent request
Review the Skillstore skill "building-chatgpt-apps" from https://skillstore.io/skills/asmayaseen-building-chatgpt-apps.md and its manifest at https://skillstore.io/api/skills/asmayaseen-building-chatgpt-apps/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.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 "building-chatgpt-apps". I need a ChatGPT App that displays a dashboard widget.
Expected outcome:
- A FastMCP server structure with a widget resource.
- The required response metadata for widget rendering.
- A local testing sequence using Developer Mode.
Using "building-chatgpt-apps". My widget does not update after I changed the HTML.
Expected outcome:
- A cache reset checklist.
- Steps to create a fresh tunnel and app registration.
- Checks for MIME type and response metadata.
Using "building-chatgpt-apps". How should widget buttons communicate with ChatGPT?
Expected outcome:
- Use simple button handlers.
- Prefer follow-up messages for reliable actions.
- Check window.openai APIs before calling them.
Security Audit
Medium Riskv11 • 7/9/2026 Open versioned report
The skill is primarily documentation and sample templates for building ChatGPT Apps, and most static findings are false positives from markdown code fences, URLs, or CSS text. The real risks are development guidance that binds a server to all interfaces, exposes it through ngrok without authentication, and uses forceful process-kill commands during cache resets.
6
Files scanned
2,270
Lines analyzed
5
Review items
0
False positives ignored
Confirmed security concerns (1)
Medium
Public Unauthenticated Development Endpoint Guidance
The skill instructs users to expose the MCP server through an ngrok tunnel and register it with Authentication set to No Auth for development. This is useful for testing but becomes risky if copied into production or used with sensitive data.
The development workflow explicitly combines a public tunnel with no authentication. The surrounding context frames it as development, so the risk is misuse rather than clear malicious intent.
Capability review items (5)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
xargs command (can execute arbitrary commands)
lsof -ti:8001 | xargs kill -9
The documentation recommends a kill pipeline that forcefully terminates processes on port 8001. It is development-oriented, but it can disrupt unrelated local processes if copied without review.
Medium
Hardcoded IP address
host="0.0.0.0",
The template binds the MCP server to 0.0.0.0, which exposes it beyond localhost when used with no added controls. This is acceptable for development only with explicit network restrictions.
Medium
xargs command (can execute arbitrary commands)
lsof -ti:8001 | xargs kill -9
The documentation recommends a kill pipeline that forcefully terminates processes on port 8001. It is development-oriented, but it can disrupt unrelated local processes if copied without review.
Medium
xargs command (can execute arbitrary commands)
lsof -ti:8001 | xargs kill -9 2>/dev/null
The documentation recommends a kill pipeline that forcefully terminates processes on port 8001. It is development-oriented, but it can disrupt unrelated local processes if copied without review.
Medium
Hardcoded IP address
host="0.0.0.0",
The template binds the MCP server to 0.0.0.0, which exposes it beyond localhost when used with no added controls. This is acceptable for development only with explicit network restrictions.
Risk Factors
⚙️ External commands (49)
references/complete_template.md:403 references/debugging.md:424 references/debugging.md:77 references/debugging.md:428 SKILL.md:15 SKILL.md:27-34 SKILL.md:34-42 SKILL.md:42-51 SKILL.md:51-59 SKILL.md:59-67 SKILL.md:67-78 SKILL.md:78-79 SKILL.md:79-85 SKILL.md:85-91 SKILL.md:91-118 SKILL.md:118-124 SKILL.md:124-142 SKILL.md:142-149 SKILL.md:149-150 SKILL.md:150-160 SKILL.md:160-217 SKILL.md:217-221 SKILL.md:221-224 SKILL.md:224-232 SKILL.md:232-237 SKILL.md:237-241 SKILL.md:241-246 SKILL.md:246-250 SKILL.md:250-367 SKILL.md:367-373 SKILL.md:373-377 SKILL.md:377-393 SKILL.md:393-397 SKILL.md:397-405 SKILL.md:405-409 SKILL.md:409-413 SKILL.md:413-416 SKILL.md:416-426 SKILL.md:426-434 SKILL.md:434-447 SKILL.md:447-448 SKILL.md:448-449 SKILL.md:449-464 SKILL.md:464-467 SKILL.md:467-471 SKILL.md:471-479 SKILL.md:479-480 SKILL.md:480-486 SKILL.md:486-488
🌐 Network access (17)
references/complete_template.md:327 references/complete_template.md:332 references/complete_template.md:336 references/debugging.md:47 references/debugging.md:50 references/debugging.md:368 references/debugging.md:388 references/debugging.md:393 references/debugging.md:442 SKILL.md:359 SKILL.md:360 SKILL.md:408 SKILL.md:415 SKILL.md:420 SKILL.md:426 SKILL.md:479 SKILL.md:363
🔑 Env variables (2)
📁 Filesystem access (2)
Audited by: claude 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/asmayaseen-building-chatgpt-apps/audits/11?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/asmayaseen-building-chatgpt-apps?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/asmayaseen-building-chatgpt-apps?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/asmayaseen-building-chatgpt-apps/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/asmayaseen-building-chatgpt-apps.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
Asmayaseen. (2026). building-chatgpt-apps security audit report (audit version 11) [Author version unspecified]. Skillstore. https://skillstore.io/skills/asmayaseen-building-chatgpt-apps/audits/11BibTeX citation
@techreport{asmayaseen-asmayaseen-building-chatgpt-apps-2026,
author = {Asmayaseen},
title = {building-chatgpt-apps security audit report (audit version 11)},
institution = {Skillstore},
year = {2026},
number = {11},
url = {https://skillstore.io/skills/asmayaseen-building-chatgpt-apps/audits/11},
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: "building-chatgpt-apps security audit report (audit version 11)"
version: "unspecified"
type: report
authors:
- name: "Asmayaseen"
date-released: "2026-07-09"
url: "https://skillstore.io/skills/asmayaseen-building-chatgpt-apps/audits/11"
identifiers:
- type: other
value: "skillstore:asmayaseen-building-chatgpt-apps:audit:11"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: High 68
Architecture
85
Maintainability
87
Content
70
Community
78
Spec Compliance
What You Can Build
Prototype a widget-based ChatGPT App
Use the templates to create a FastMCP app that returns embedded widget HTML and structured content.
Debug ChatGPT widget rendering
Follow the troubleshooting steps for metadata, MIME type, caching, and tunnel connectivity problems.
Teach Apps SDK architecture
Use the diagrams and examples to explain how ChatGPT, widgets, and MCP servers work together.
Try These Prompts
Create a basic app plan
Use this skill to outline a simple ChatGPT App with one FastMCP tool and one embedded widget.
Generate widget metadata
Use this skill to show the response metadata I need for a ChatGPT widget named dashboard-widget.
Troubleshoot loading failures
Use this skill to diagnose why my ChatGPT App widget is loading forever and list checks in priority order.
Review an app template for production readiness
Use this skill to review my ChatGPT App template and identify changes needed before production deployment.
Best Practices
- Use the templates as development starting points and add authentication before public deployment.
- Test widgets inside ChatGPT because sandbox behavior differs from a normal browser.
- Keep widget HTML self-contained with inline styles and explicit metadata.
Avoid
- Do not publish a public tunnel with no authentication for real user data.
- Do not assume complex browser JavaScript will work reliably inside the widget sandbox.
- Do not reuse cached widget URIs when testing major UI changes.
Frequently Asked Questions
What does this skill help build?
It helps build ChatGPT Apps that combine MCP tools with embedded HTML widgets.
Does it cover standard MCP servers?
It focuses on MCP servers that support ChatGPT widgets. Standard MCP-only servers are outside its main scope.
Why does it mention ngrok?
ChatGPT needs a reachable HTTPS endpoint during development, and ngrok is used as a local tunnel example.
Can I use the templates in production?
Use them as a starting point only. Add authentication, deployment controls, and safer process management first.
What is the most important widget metadata?
Tool responses must include the openai.com/widget metadata with an embedded widget resource.
Why should I test inside ChatGPT?
The widget sandbox provides window.openai APIs and behavior that normal browser previews do not reproduce.
Developer Details
Author
AsmayaseenLicense
MIT
Skillstore revision
r1
Version notice
The author did not declare a version.
Ref
1ffa7643651792ccb4bd3b15d924d2c97edff755
Maintenance freshness
7/18/2026
Usage
7 downloads · 195 views
File structure