Background
A recurring theme in this IG is that clients have been slow to implement support for resources. Before recommending skills-as-resources as a convention, we need to know which clients actually support resources/list, resources/read, resources/subscribe, and resource templates.
"Clients have been slow to implement support for resources. Had some parallel primitive 'skills' been implemented, I'm not sure clients would have implemented them any faster. Basically they all went for 'tools' and have slowly been getting around to implementing other primitives." — Cliff Hall
Scope
Survey the following MCP clients and document their resource support:
| Client |
resources/list |
resources/read |
resources/subscribe |
Resource Templates |
Model can access? |
Notes |
| Claude Desktop |
? |
? |
? |
? |
? |
|
| Claude Code |
? |
? |
? |
? |
? |
|
| VS Code (Copilot) |
? |
? |
? |
? |
? |
|
| Cursor |
? |
? |
? |
? |
? |
|
| Windsurf |
? |
? |
? |
? |
? |
|
| Zed |
? |
? |
? |
? |
? |
|
| Continue |
? |
? |
? |
? |
? |
|
For each client, note:
- Whether resources are auto-loaded or require user action
- Whether the model can interact with resources directly (or if it's application-controlled only)
- Any known limitations or quirks
Acceptance Criteria
References
Background
A recurring theme in this IG is that clients have been slow to implement support for resources. Before recommending skills-as-resources as a convention, we need to know which clients actually support
resources/list,resources/read,resources/subscribe, and resource templates.Scope
Survey the following MCP clients and document their resource support:
For each client, note:
Acceptance Criteria
docs/experimental-findings.mdas a new sectionReferences