Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+21 moreClaude CodeCursorCodex CLIMuseDots
Manus
Grok
Grok Bot+18 agentsDropCheck
Simulate 100 loot drops to identify broken weights and generate balanced CSV weight files.
Free
DropCheck
Example session with this skill installed
Drop Check for this table
Common Sword, 70, 5g
Iron Shield, 25, 15g
Golden Idol, 4, 500g
Dragon Tooth, 1, 1000g
- Read your context and instructions
- Compiled the dropcheck
- Generated the data export
I have processed the loot table simulation using a 100-roll distribution. The results show the Common Sword is currently eating the table, while the Dragon Tooth remained a ghost item during this seed.
DROP CHECK
Seed: 1
Rolls: 100
PILE
| Item | Rarity | Old weight | Hits | Hit % | Gold | Fixed weight |
|---|---|---|---|---|---|---|
| Common Sword | common | 70 | 78 | 78% | 390g | 55 |
| Iron Shield | common | 25 | 20 | 20% | 300g | 30 |
| Golden Idol | common | 4 | 2 | 2% | 1000g | 12 |
| Dragon Tooth | common | 1 | 0 | 0% | 0g | 3 |
EATER: Common Sword
GHOST: Dragon Tooth
JUNK GOLD: 40.8% of dropped value
DRY STREAK: 100 kills without Dragon Tooth
FIX
- Reduced Common Sword weight to prevent table saturation.
- Buffered Golden Idol and Dragon Tooth weights to ensure visibility in small sample sizes.
- Kept every name. Did not add items.
Files generated: drop-pile.md, weights.csv
name,old_weight,hits,fixed_weight
Common Sword,70,78,55
Iron Shield,25,20,30
Golden Idol,4,2,12
Dragon Tooth,1,0,3
Next steps
- Review the
weights.csvfile to ensure the new distribution fits your intended rarity curve. - Run the check again with a different seed (e.g.,
seed: 42) to verify the Rare drop consistency. - Import the
fixed_weightcolumn into your game's master item database.
weights.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Loot tables in game development often break when a single high-weight item suppresses the rest of the pool, or when rare items simply never appear due to math errors. This skill identifies "table eaters" and "ghost items" by simulating 100 drops and providing a corrected weight file instantly.
What it does
- Simulate drops by running 100 rolls using a dedicated Python script to ensure mathematical accuracy.
- Identify table eaters by flagging items whose weights disproportionately consume the probability pool.
- Detect ghost items by listing every item that failed to drop a single time during the simulation.
- Calculate junk value by analyzing the gold percentage represented by common or low-value drops.
- Generate fixed weights by producing a balanced
weights.csvthat maintains your original item names.
How it works
- Input table by pasting your item names, weights, and optional rarities into the chat.
- Run simulation using the
roll.pyscript to generate a JSON-based distribution report. - Review report to see hit percentages, dry streaks for rares, and gold distribution.
- Export fixes by saving the provided
drop-pile.mdandweights.csvfiles to your project.
Frameworks & tools
This skill requires a Python 3 environment to execute the roll.py simulation script. It processes JSON input and outputs structured Markdown and CSV files.
Why this beats prompting it yourself
Standard LLMs are notorious for "hallucinating" math and distribution percentages. This skill uses a deterministic script to perform the rolls, ensuring the hit counts and fixed weights are calculated via code rather than prose estimation.
Use cases
- Debugging why a specific "Rare" item hasn't dropped for players in weeks.
- Balancing economy by checking the "Junk Gold" percentage of a loot table.
- Stress-testing new loot pools before committing them to a database.
Known limitations
Requires at least two unique item names to run. It does not support affixes, pity systems, or multiple nested tables in a single pass.
Requires python3 to run the roll script. Limited to 100 rolls per simulation. Does not handle nested tables, affixes, or pity mechanics.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
4 uses
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 29 days ago
- Free with an Agensi account
Permissions
No special permissions declared or detected