---
name: check-my-credits
description: Use this skill when the user asks about their credit balance, usage, or tool costs — or at the start of any session involving significant enrichment work.
---

# Check My Credits

Retrieve account balance, per-tool costs, and usage history. All calls are free.

## Triggers:

"how many credits do I have", "check my balance", "how much have I spent", "what does [tool] cost", "will I have enough credits for this", "show me my usage", "am I running low". Both tools are free and never charge credits.

## Always run in parallel

- `MoltSets:get_billing` — plan, credit balance, renewal date, credit cost per tool
- `MoltSets:get_usage` with `period: billing_cycle` — credits consumed this cycle with per-day breakdown

## Standard output

```
Plan:     Pro
Balance:  4,280 credits
Renewal:  2025-02-01

Usage this billing cycle: 1,720 credits

Credit costs per tool:
  linkedin_slug_search .............. X cr
  enrich_email ...................... X cr
  enrich_personal_email ............. X cr
  enrich_phone ...................... X cr
  linkedin_to_business_profile ...... X cr
  hem_to_best_linkedin .............. X cr
  ip_to_company ..................... X cr
  ip_to_hem ......................... X cr
  (show all tools from get_billing)
```

## Cost estimation

When the user is about to run a batch operation, calculate before proceeding:

```
Estimated: 50 profiles × X credits (enrich_email) = XXX credits
Balance:   4,280 credits
After:     ~3,980 credits
```

**If estimated cost > balance:** stop. Report the shortfall, how many rows/calls the balance can cover, and ask whether to proceed partially or wait for a top-up.

## Usage period queries

| User asks | Use period |
|---|---|
| "today" | `today` |
| "this week" | `week` |
| "this month" | `month` |
| "this billing cycle" / default | `billing_cycle` |

## Proactive low-balance warning

If balance < 500 credits, warn proactively before any enrichment operation — even if the user didn't ask to check.

## Key rule: "not found" is free

Results with `status: "not_found"` are never charged. Factor this into cost estimates for data with expected low match rates (personal email ~25–40%, phone ~30–50%).