Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Automate Gumroad product management, sales tracking, license verification, and webhook subscriptions using natural language through the Composio MCP integration.
.claude/skills/composiohq-gumroad-automation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 48% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 260% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -1% | 0% |
Automate your Gumroad storefront -- list products, track sales, verify licenses, and manage real-time webhooks -- all through natural language commands.
Toolkit docs: composio.dev/toolkits/gumroad
https://rube.app/mcp
Retrieve every product in your authenticated Gumroad account to get product IDs for downstream operations.
Tool: GUMROAD_LIST_PRODUCTS
Example prompt: > "List all my Gumroad products"
Parameters: None required -- returns all products for the authenticated account.
Retrieve successful sales with optional filtering by email, date range, product, or pagination.
Tool: GUMROAD_GET_SALES
Example prompt: > "Show me all Gumroad sales from January 2025 for product prod_ABC123"
Key parameters:
after -- ISO8601 date/time to filter sales after (e.g., 2025-01-01T00:00:00Z)before -- ISO8601 date/time to filter sales beforeemail -- Filter by customer email addressproduct_id -- Filter by specific product IDpage -- Page number for paginated results (minimum 1)Check if a license key is valid against a specific product, inspect usage count, or verify membership entitlement.
Tool: GUMROAD_VERIFY_LICENSE
Example prompt: > "Verify license key ABCD-EFGH-IJKL-MNOP for product prod_ABC123"
Key parameters (all required):
product_id -- The product ID to verify against (required for products created on/after Jan 9, 2023)license_key -- The license key string (e.g., ABCD-EFGH-IJKL-MNOP)increment_uses_count -- Whether to increment usage count (defaults to true)Set up real-time event notifications by subscribing your endpoint URL to specific Gumroad resource events.
Tool: GUMROAD_SUBSCRIBE_TO_RESOURCE
Example prompt: > "Subscribe my webhook https://example.com/hook to Gumroad sale events"
Key parameters (all required):
resource_name -- One of: sale, refund, dispute, dispute_won, cancellation, subscription_updated, subscription_ended, subscription_restartedpost_url -- Your endpoint URL that receives HTTP POST notificationsReview existing webhook subscriptions for a given resource type before adding new ones to avoid duplicates.
Tool: GUMROAD_GET_RESOURCE_SUBSCRIPTIONS
Example prompt: > "Show all my active Gumroad webhook subscriptions for sale events"
Key parameters (required):
resource_name -- One of the eight supported event types (e.g., sale, refund)product_id parameter. Older products may work without it but providing it is recommended.page parameter.GUMROAD_GET_RESOURCE_SUBSCRIPTIONS to check for existing subscriptions and avoid duplicate webhooks.2025-01-01T00:00:00Z), not plain dates.| Action | Tool Slug | Required Params | |---|---|---| | List products | GUMROAD_LIST_PRODUCTS | None | | Get sales | GUMROAD_GET_SALES | None (all optional filters) | | Verify license | GUMROAD_VERIFY_LICENSE | product_id, license_key | | Subscribe to events | GUMROAD_SUBSCRIBE_TO_RESOURCE | resource_name, post_url | | List webhook subs | GUMROAD_GET_RESOURCE_SUBSCRIPTIONS | resource_name |
Powered by Composio
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,661 | 6,489 | -3% | 1 | 1 | 0% | 1,290 | 1,540 | +19% | 0 | 0 | — |
case-02 | fail→fail | 2,665 | 7,059 | +165% | 1 | 1 | 0% | 473 | 1,649 | +249% | 0 | 0 | — |
case-03 | fail→fail | 3,045 | 2,537 | -17% | 1 | 1 | 0% | 377 | 1,404 | +272% | 0 | 0 | — |
case-04 | fail→pass | 9,055 | 7,812 | -14% | 1 | 1 | 0% | 1,704 | 2,525 | +48% | 0 | 0 | — |
case-05 | fail→pass | 4,223 | 6,551 | +55% | 1 | 1 | 0% | 655 | 2,356 | +260% | 0 | 0 | — |
case-10 | fail→pass | 9,337 | 2,647 | -72% | 1 | 1 | 0% | 1,755 | 1,556 | -11% | 0 | 0 | — |
case-06 | pass→pass | 3,859 | 4,555 | +18% | 1 | 1 | 0% | 668 | 1,923 | +188% | 0 | 0 | — |
case-07 | fail→pass | 5,054 | 1,942 | -62% | 1 | 1 | 0% | 866 | 1,396 | +61% | 0 | 0 | — |
case-08 | fail→pass | 10,015 | 3,647 | -64% | 1 | 1 | 0% | 1,714 | 1,690 | -1% | 0 | 0 | — |
case-09 | fail→pass | 8,693 | 2,788 | -68% | 1 | 1 | 0% | 1,716 | 1,573 | -8% | 0 | 0 | — |
case-11 | pass→pass | 9,346 | 2,435 | -74% | 1 | 1 | 0% | 1,853 | 1,621 | -13% | 0 | 0 | — |
case-12 | fail→pass | 9,564 | 3,221 | -66% | 1 | 1 | 0% | 1,576 | 1,731 | +10% | 0 | 0 | — |
case-13 | fail→pass | 7,516 | 2,903 | -61% | 1 | 1 | 0% | 1,230 | 1,551 | +26% | 0 | 0 | — |
case-14 | fail→pass | 9,826 | 2,158 | -78% | 1 | 1 | 0% | 1,773 | 1,372 | -23% | 0 | 0 | — |
case-15 | pass→pass | 3,832 | 2,578 | -33% | 1 | 1 | 0% | 587 | 1,530 | +161% | 0 | 0 | — |
case-16 | fail→pass | 7,650 | 3,509 | -54% | 1 | 1 | 0% | 1,410 | 1,644 | +17% | 0 | 0 | — |
case-17 | fail→pass | 14,169 | 3,285 | -77% | 1 | 1 | 0% | 2,550 | 1,694 | -34% | 0 | 0 | — |
case-18 | pass→pass | 5,247 | 4,237 | -19% | 1 | 1 | 0% | 939 | 1,831 | +95% | 0 | 0 | — |
case-19 | fail→pass | 6,311 | 3,104 | -51% | 1 | 1 | 0% | 1,098 | 1,631 | +49% | 0 | 0 | — |
case-20 | fail→pass | 8,649 | 3,227 | -63% | 1 | 1 | 0% | 1,180 | 1,665 | +41% | 0 | 0 | — |
case-21 | fail→pass | 7,682 | 4,687 | -39% | 1 | 1 | 0% | 1,580 | 1,854 | +17% | 0 | 0 | — |
case-22 | pass→pass | 5,436 | 1,888 | -65% | 1 | 1 | 0% | 797 | 1,412 | +77% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 20 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +64 percentage points is the difference between those two pass rates over the 20 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.