Skip to content
Back to search

chatserp

MarketplaceWorks with Claude CodeGitHub

ChatSERP: get your product inside ChatGPT. Builds, tests and submits a ChatGPT plugin (Apps SDK / MCP) for any product.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/oreldm--chatserp/badge.svg)](/listings/oreldm--chatserp)

HTML

<a href="/listings/oreldm--chatserp"><img src="/api/listings/oreldm--chatserp/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

In Claude Code (one step):

/plugin install chatserp --marketplace Oreldm/chatserp

Or add the marketplace, then install:

/plugin marketplace add Oreldm/chatserp
/plugin install chatserp@chatserp

Or copy the skill by hand:

git clone https://github.com/Oreldm/chatserp
cp -r chatserp/skills/chatserp ~/.claude/skills/        # macOS / Linux
# Windows: xcopy /E /I chatserp\skills\chatserp %USERPROFILE%\.claude\skills\chatserp

Then in Claude Code: *"Make a ChatGPT plugin for my product"*, or `/chatserp`.

Test any MCP endpoint without the skill:

# check_mcp.py = the code block in skills/chatserp/tools/check_mcp.md
python check_mcp.py https://your.site/mcp \
  --call search_items '{"query":"pricing"}' \
  --expect-error search_items '{"query":""}'

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 13 lines
  1. {
  2. "name": "chatserp",
  3. "description": "ChatSERP: get your product inside ChatGPT. Builds, tests and submits a ChatGPT plugin (Apps SDK / MCP) for any product.",
  4. "owner": { "name": "Oreldm", "url": "https://github.com/Oreldm" },
  5. "plugins": [
  6. {
  7. "name": "chatserp",
  8. "source": "./",
  9. "description": "Turn any product, website or API into a ChatGPT plugin that passes OpenAI review: go/no-go value test, real-data tools, MCP server template, hosting, live endpoint check, submission pack.",
  10. "category": "productivity"
  11. }
  12. ]
  13. }
README.md · README · 117 lines
  1. <p align="center"><img src="assets/banner.png" alt="ChatSERP" width="100%"></p>
  2. <p align="center">
  3. <b>Get your product inside ChatGPT, where people already ask for it.</b><br>
  4. A <a href="https://docs.claude.com/en/docs/claude-code/skills">Claude Code skill</a> that builds, tests and ships a ChatGPT plugin for any product, website or API.
  5. </p>
  6. <p align="center">
  7. <img alt="license" src="https://img.shields.io/badge/license-MIT-3fb950">
  8. <img alt="Claude Code skill" src="https://img.shields.io/badge/Claude%20Code-skill-d97757">
  9. <img alt="MCP" src="https://img.shields.io/badge/MCP-Apps%20SDK-58a6ff">
  10. <img alt="node" src="https://img.shields.io/badge/node-22%2B-8b949e">
  11. </p>
  12. ---
  13. ChatGPT now runs third-party **plugins** (built on the Apps SDK / MCP) right inside the conversation.
  14. Someone asks *"convert this PDF to Excel"*, *"what's the best move here?"* or *"how do I prep for this team?"*,
  15. and the plugin that actually answers it gets the user.
  16. **Every big platform has a short window where getting users is cheap:** Facebook apps in 2007, the App Store in 2008, TikTok in 2019. ChatGPT plugins are that window right now.
  17. ChatSERP is the playbook as one skill. Point Claude Code at your product and it runs the whole thing:
  18. ```
  19. /chatserp my-product
  20. ```
  21. ## What it does
  22. | Step | What happens |
  23. |---|---|
  24. | **1. Go / no-go** | One-line value test. Kills ideas OpenAI will reject (ad-only plugins, demos, link-only "apps") *before* you write code. |
  25. | **2. Tool design** | 2-4 tools built on **your real data or engine**, with metadata written the way ChatGPT matches tools to prompts ("Use this when..."). |
  26. | **3. MCP server** | Drop-in Node 22 template: stateless Streamable HTTP `/mcp`, zod validation, rate limit, optional inline widget, domain-check route. |
  27. | **4. Hosting** | Cheapest path on your own domain: Firebase Functions, Cloud Run or Vercel. |
  28. | **5. Live check** | `check_mcp.py` runs initialize, tools/list and tools/call against the real URL, and fails on bad metadata or empty answers. |
  29. | **6. Submission pack** | Privacy section, listing copy, 5 positive and 3 negative test prompts, release notes, plugin zip, portal steps. |
  30. | **7. Submit** | Walks the OpenAI Plugins portal with you. You keep the clicks that are yours: identity check and accepting the terms. |
  31. ## Install
  32. In Claude Code (one step):
  33. ```
  34. /plugin install chatserp --marketplace Oreldm/chatserp
  35. ```
  36. Or add the marketplace, then install:
  37. ```
  38. /plugin marketplace add Oreldm/chatserp
  39. /plugin install chatserp@chatserp
  40. ```
  41. Or copy the skill by hand:
  42. ```bash
  43. git clone https://github.com/Oreldm/chatserp
  44. cp -r chatserp/skills/chatserp ~/.claude/skills/ # macOS / Linux
  45. # Windows: xcopy /E /I chatserp\skills\chatserp %USERPROFILE%\.claude\skills\chatserp
  46. ```
  47. Then in Claude Code: *"Make a ChatGPT plugin for my product"*, or `/chatserp`.
  48. Test any MCP endpoint without the skill:
  49. ```bash
  50. # check_mcp.py = the code block in skills/chatserp/tools/check_mcp.md
  51. python check_mcp.py https://your.site/mcp \
  52. --call search_items '{"query":"pricing"}' \
  53. --expect-error search_items '{"query":""}'
  54. ```
  55. ## The honest part (read before you build)
  56. - **ChatGPT will not start recommending you on day one.** New plugins are found by name search or a direct link. Suggestions in conversation and directory placement are picked by OpenAI, and you can't apply for them. Real usefulness and sharp metadata are the levers, and that's what this skill optimizes.
  57. - **Ad-only plugins get rejected.** "Download our app" is not a plugin. The skill's go / no-go step exists for exactly this.
  58. - **No selling digital goods inside the plugin.** Commerce is physical goods only, with no checkout links. Use the plugin to bring users to you, not to make the sale.
  59. - **Your developer account must be verified** (platform.openai.com > Settings > Organization). That's an identity check only you can do.
  60. The skill reads OpenAI's live docs on every run, because the rules move.
  61. ## Services the skill talks to
  62. ChatSERP has no server of its own and collects nothing (see [PRIVACY.md](PRIVACY.md)). While it works, Claude may reach:
  63. - **developers.openai.com**: reads OpenAI's public plugin docs (no user data is sent).
  64. - **Your own hosting** (Firebase, Cloud Run, Vercel...): deploys your plugin server with your own CLI login.
  65. - **Your own plugin endpoint**: `check_mcp.py` sends test calls to the URL you give it.
  66. - **platform.openai.com**: the plugin submission, in your own OpenAI account, with you at the keyboard.
  67. ## Good fits
  68. - A **real engine:** calculators, converters, solvers, analyzers, game move engines
  69. - **Your own data:** catalogs, knowledge bases, findings, listings, docs
  70. - The **user's account** (OAuth): their orders, files, stats
  71. - **Interactive widgets:** a board, a map, a chart, a playable mini-game
  72. ## Repo layout
  73. ```
  74. skills/chatserp/
  75. SKILL.md the playbook Claude follows
  76. templates/server.md MCP server template (code block -> server.mjs) (Node 22, @modelcontextprotocol/sdk)
  77. tools/check_mcp.md live endpoint checker (code block -> check_mcp.py) (Python stdlib only)
  78. ```
  79. ## Contributing
  80. Found a new review rule, a rejection reason, or a hosting recipe? Open a PR. Rejection notes are gold.
  81. ---
  82. <sub>Not affiliated with, endorsed by, or sponsored by OpenAI or Anthropic. "ChatGPT" is a trademark of OpenAI. MIT licensed.</sub>
  83. If ChatSERP got your plugin approved, star the repo so the next builder finds it.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.