Skip to content

improve: fix docs skill title, clarify mcp.json placeholders, add installation steps to README#2

Open
yaron-reichert-cld wants to merge 1 commit intocloudinary-devs:mainfrom
yaron-reichert-cld:improve/onboarding-and-ux
Open

improve: fix docs skill title, clarify mcp.json placeholders, add installation steps to README#2
yaron-reichert-cld wants to merge 1 commit intocloudinary-devs:mainfrom
yaron-reichert-cld:improve/onboarding-and-ux

Conversation

@yaron-reichert-cld
Copy link
Contributor

Summary

Three usability issues found during testing that would confuse new users or the agent itself.

Fix 1 — cloudinary-docs skill: remove # My Skill placeholder title

skills/cloudinary-docs/SKILL.md had a leftover template heading # My Skill instead of a real title. Changed to # Cloudinary Documentation Skill. Small but it looks unprofessional and could confuse an agent that reads the skill name vs its H1.

Fix 2 — mcp.json: credential placeholders look like real values

Before:

"cld-cloud-name": "cloud_name",
"cld-api-key": "api_key",
"cld-secret": "api_secret"

After:

"cld-cloud-name": "YOUR_CLOUD_NAME",
"cld-api-key": "YOUR_API_KEY",
"cld-secret": "YOUR_API_SECRET"

The original values ("cloud_name", "api_key", "api_secret") read like real credentials rather than placeholders. New users would likely paste this file and be confused why MediaFlows doesn't authenticate.

Fix 3 — README: add actual installation instructions and example prompts

Before: Step 1 simply said "Install this plugin in Cursor/Claude Code" — no guidance on how.

After: Added step-by-step instructions for both Cursor (copy mcp.json, copy skills/, create rules) and Claude Code (--plugin-dir). Also added 5 example prompts to help users understand what they can ask immediately after installing.

How this was found

Identified during a test run of the plugin, focusing on the first-time user experience and agent configuration quality.

Made with Cursor

…tall steps to README

Three usability issues identified during plugin testing:

1. cloudinary-docs/SKILL.md had a "# My Skill" placeholder title left
   over from the skill template — replaced with the correct title.

2. mcp.json credential placeholders ("cloud_name", "api_key",
   "api_secret") looked like real values rather than placeholders.
   Renamed to "YOUR_CLOUD_NAME" / "YOUR_API_KEY" / "YOUR_API_SECRET"
   to make it obvious these must be replaced.

3. README "Getting Started" step 1 said "Install this plugin in
   Cursor/Claude Code" with no instructions on how. Added step-by-step
   installation instructions for both Cursor and Claude Code, plus
   five example prompts to help users get started quickly.

Made-with: Cursor
Comment on lines +32 to +46
### Cursor

1. Copy the contents of `mcp.json` into your project's `.cursor/mcp.json` (create the file if it doesn't exist).
2. Copy the `skills/` directory into your project root.
3. In `.cursor/rules/`, create one rule file per skill — each pointing to its `SKILL.md`. Cursor will use the rule description to decide when to invoke the skill.
4. For `cloudinary-mediaflows`, replace `YOUR_CLOUD_NAME`, `YOUR_API_KEY`, and `YOUR_API_SECRET` in `mcp.json` with your Cloudinary credentials (found in the [Cloudinary Console](https://console.cloudinary.com/settings/api-keys)).
5. On first use, approve the OAuth login prompt that Cursor shows for the other four MCP servers.

### Claude Code

1. Run: `claude --plugin-dir /path/to/cloudinary-plugin`
2. For `cloudinary-mediaflows`, replace the credential placeholders in `mcp.json` before starting.
3. On first use, complete the OAuth login for the other four MCP servers.

Once installed, ask your agent anything about Cloudinary — uploads, transformations, metadata, analysis, and more. The agent will automatically use the right skill and MCP server for the job.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This entire section is incorrect. Both Cursor/Claude have a marketplace included, and the installation is built in inside.

# Cloudinary Documentation Skill

Helps developers to integrate Cloudinary into their applications by providing documentation and code examples retrieved directly from the optimized .md files in the Cloudinary documentation rather than having parse the HTML pages.
Helps developers integrate Cloudinary into their applications by providing documentation and code examples retrieved directly from the optimized .md files in the Cloudinary documentation, rather than having to parse HTML pages.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually the original is an awful description all together (bad work on Cursor's part lol being too honest).

Let's change to:

Suggested change
Helps developers integrate Cloudinary into their applications by providing documentation and code examples retrieved directly from the optimized .md files in the Cloudinary documentation, rather than having to parse HTML pages.
Helps developers integrate Cloudinary into their applications by providing documentation and code examples.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@sveta-slepner - Do we want to be a bit more specific on what it does than just "providing"? - i.e.

Ensures that the relevant pages from the latest version of the Cloudinary documentation, including code examples from all supported SDKs, are included in the context whenever asking any question relating to Cloudinary

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Feel free to change to whatever you see fit. It was generated by Cursor :)

@jackieros jackieros assigned carlevison and unassigned carlevison Mar 12, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants