Token Management
Create and manage MCP tokens for secure AI integration.
Token Management
MCP tokens provide secure access to Edsger's MCP API for AI agents, the CLI, and external tools.
Creating a Token
- Navigate to "MCP" in the sidebar
- Click "Create Token"
- Configure the token:
- Name: Descriptive name for the token
- Permissions: Select allowed resources
- Rate Limits: Set request limits
Token Permissions
Configure which resources the token can access:
- Products: Read/write product data
- Features: Manage features and their lifecycle
- User Stories: Create, update, and delete stories
- Test Cases: Manage test cases and their status
- Test Reports: Create and update test reports
- Checklists: Manage checklists and checklist items
Rate Limiting
Protect your resources with rate limits:
| Setting | Description | Default |
|---|---|---|
| Per Minute | Requests per minute | 60 |
| Per Hour | Requests per hour | 1000 |
| Max Context | Maximum response size | 10000 |
Rate limits are enforced at the token level. Exceeding limits returns an error response.
Token Authentication
Tokens are used as bearer tokens in the Authorization header:
Authorization: Bearer YOUR_MCP_TOKENTokens are validated via the validate_mcp_token RPC function, which checks validity, permissions, and rate limits in a single call.
Security Best Practices
Token Storage
- Never commit tokens to version control
- Use environment variables (e.g.,
EDSGER_MCP_TOKEN) - Rotate tokens regularly
Minimal Permissions
- Only grant necessary permissions
- Create separate tokens for different use cases (CLI vs. external integrations)
- Revoke unused tokens
Monitoring
- Review access logs regularly (all requests logged to
mcp_access_logs) - Track usage statistics via
mcp_usage_stats - Set up alerts for unusual activity
Revoking Tokens
To revoke a token:
- Go to MCP settings
- Find the token
- Click "Revoke"
Revoked tokens are immediately invalidated. All subsequent requests with the token will be rejected.
Usage Analytics
Track token usage in the MCP dashboard:
- Request count
- Error rate
- Most accessed resources
- Response times