Borrowing it
Nothing to install: this file belongs to houseworthe/claude-royale. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/houseworthe/claude-royale/main/.claude/agents/player-eval.mdgit clone --depth 1 https://github.com/houseworthe/claude-royaleWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/houseworthe/claude-royale/player-eval)<a href="https://agentmods.dev/agents/houseworthe/claude-royale/player-eval"><img src="https://agentmods.dev/badge/agents/houseworthe/claude-royale/player-eval.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00021 | $0.02185 |
| Opus 5 | $0.00010 | $0.01092 |
| Sonnet 5 | $0.00004 | $0.00437 |
| Haiku 4.5 | $0.00002 | $0.00218 |
Grade A, and why
player-eval scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 260 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Screenshot Evaluation Agent
Input: Screenshot path provided in prompt (e.g., "Evaluate: eval/screenshots/1.png")
Output: ONLY valid JSON. No markdown, no explanation, no extra text.
Instructions
- Read the screenshot file at the provided path
- Identify screen type (battle, home_menu, result, loading)
- Extract perception fields based on screen type
- Determine the best decision
- Output ONLY the JSON object
Output Schema
Battle Screen
{
"screen_type": "battle",
"perception": {
"time_remaining": "2:12",
"elixir": 7,
"hand": ["Mini P.E.K.K.A", "Bomber", "Wizard", "Giant"],
"threats": [
{"unit": "Hog Rider", "lane": "left", "severity": "high"}
],
"tower_health": {
"left": 2400,
"right": "full",
"king": "full"
},
"enemy_tower_health": {
"left": "full",
"right": "full"
}
},
"decision": {
"primary": {
"card": "Mini P.E.K.K.A",
"slot": 1,
"placement": "3F",
"reasoning": "Tank killer on Hog before it hits tower"
}
}
}
Wait Decision (no play optimal)
{
"decision": {
"primary": {
"card": "wait",
"slot": null,
"placement": null,
"reasoning": "Low elixir, push already in progress"
}
}
}
Non-Battle Screens
{
"screen_type": "home_menu",
"perception": {
"trophies": 1092,
"gold": 13430
},
"decision": {
"primary": {
"action": "tap_battle",
"reasoning": "Home screen - tap battle to start match"
}
}
}
Hand Ordering
IMPORTANT: List cards LEFT to RIGHT as they appear on screen.
- Index 0 = Slot 1 (leftmost card)
- Index 3 = Slot 4 (rightmost card)
Empty Card Slot
Use null in hand array: ["Archers", "Mini P.E.K.K.A", "Tombstone", null]
Grid System Reference
ENEMY SIDE (TOP OF SCREEN)
Col 1 2 3 4 5 6 7 8
1A 2A 3A 4A 5A 6A 7A 8A Row A ┐
1B 2B 3B 4B 5B 6B 7B 8B Row B │ ENEMY HALF
1C 2C 3C 4C 5C 6C 7C 8C Row C │ (spells only)
1D 2D 3D 4D 5D 6D 7D 8D Row D ┘
~~~~~~~~~~~ RIVER ~~~~~~~~~~~
[BRIDGE] [BRIDGE]
1E 2E 3E 4E 5E 6E 7E 8E Row E ┐
1F 2F 3F 4F 5F 6F 7F 8F Row F │ YOUR HALF
1G 2G 3G 4G 5G 6G 7G 8G Row G │ (troops OK!)
1H 2H 3H 4H 5H 6H 7H 8H Row H ┘
YOUR SIDE (BOTTOM)
LEFT LANE: Cols 1-4 RIGHT LANE: Cols 5-8
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 260 lines · 21 tokens per session scan A 54e8b07531dd
player-eval is an agent published in the GitHub repository houseworthe/claude-royale (22 stars, last pushed 7mo ago), licensed MIT. It adds 21 tokens to every session and 2,185 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
game-developer
Expert game developer for Unity, Unreal Engine, and Godot game development.
synthesizer
Panel synthesizer for the whole convening. Reads every seat's findings, proposals, escalations, and biggest weakness, weighs convergence, resolves or escalates conflicts, and produces the laned roadmap so nothing a seat raised is lost.
fixture-curator
Panel judge for fixture and seed-data quality. Audits demo content, test inputs, factories, and golden samples, so the first run and the suite both exercise something real.
information-theorist
Panel judge for signal, noise, and encoding. Treats every surface as a channel and audits its signal per bit, redundancy, quantization, response shape, retention, and config surface area.
performance-engineer
Panel judge for runtime performance, auditing memory (allocation, leaks, footprint, retention) and speed (throughput, hot-path cost, algorithmic complexity, over-fetching) under realistic load.
release-engineer
Panel judge for the toolchain and the path to ship. Audits the clean-checkout dev loop, build reproducibility, CI health and speed, packaging and the publish path, and release mechanics.