-
Notifications
You must be signed in to change notification settings - Fork 127
Closed
Description
When using the claude-supermemory plugin (v0.0.2) and making API calls to Supermemory (e.g., search-memory.cjs, add-memory.cjs), the 403 error response incorrectly references "The Clawdbot plugin" instead of "The Claude Code plugin":
{
"error": "The Clawdbot plugin requires a Pro plan or higher. Upgrade at https://supermemory.ai/pricing"
}Expected Behavior
The error message should reference the correct product:
{
"error": "The Claude Code plugin requires a Pro plan or higher. Upgrade at https://supermemory.ai/pricing"
}Steps to Reproduce
- Install the plugin:
/plugin marketplace add supermemoryai/claude-supermemory /plugin install claude-supermemory
- Configure a valid API key (format:
sm_*) - Run any memory operation (search or save)
- Observe the 403 error referencing "Clawdbot plugin"
Environment
| Field | Value |
|---|---|
| Plugin version | 0.0.2 |
| Claude Code | Latest |
| OS | Linux (Ubuntu) |
| Node.js | v22 |
Concerns
1. Misleading Error Message
Users of claude-supermemory have no context about what "Clawdbot" is. The error should clearly reference "Claude Code plugin" to avoid confusion.
2. Potential Data Privacy / Security Concern
The fact that the Claude Code plugin API returns a "Clawdbot plugin" error suggests both integrations may share the same backend API endpoint and/or authentication pathway. This raises questions:
- Are Claude Code plugin users' data and Clawdbot users' data isolated? If both plugins hit the same API with the same key validation, is there proper data segregation (separate container tags, separate storage)?
- Could API keys from one integration inadvertently grant access to data from the other? If a user has both integrations configured, are memories properly scoped?
- Is the API key validation path shared? The identical error response suggests the backend doesn't distinguish between the two integrations, which could mean user context/metadata leaks between products.
3. Transparency Request
It would be helpful if Supermemory clarified in the documentation:
- Whether Claude Code and Clawdbot share backend infrastructure
- How user data isolation is ensured between the two products
- Whether API keys are scoped per-integration or shared
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels