GregorStocks/mage-bench
Magic: The Gathering harness for LLMs
First Claude commit: Mar 15, 2026Last Claude commit: 2mo agoDiscovered: Mar 16, 2026
Recent Claude Commits
File refactoring issues for oversized modules (#1293)
171e3922mo agomessage_footerRemove legacy skeleton player type (#1289)
a65bbcb2mo agomessage_footerRemove potato and staller bridge modes (#1284)
ecb619d2mo agomessage_footerFix observer watch startup race (#1285)
b58f7532mo agomessage_footerFix code scanning alerts in export loading and chat decoding (#1283)
fff1b182mo agomessage_footerSimplify round-robin matchmaker (#1282)
79117242mo agomessage_footerLaunch season 2 and harden startup/tool failures (#1281)
8789a0c2mo agomessage_footerFix observer watch attach flake (#1280)
5a306e12mo agomessage_footerMake golden tests use fresh clients per case (#1279)
e3d3d472mo agomessage_footerAdd architecture diagram to docs (#1273)
3fa59782mo agomessage_footerFail fast on bridge startup config and MCP tool checks (#1276)
19d82b12mo agomessage_footerAdd new models, retire Gemini 3 Pro, remove inactive toggle (#1268)
8ae725e2mo agomessage_footerFix blunder annotation misattribution of target selection (#1260)
d3528e32mo agomessage_footerMigrate snapshot export records to dataclasses (#1249)
ceee1f82mo agomessage_footerRetire the legacy analysis decision compatibility layer (#1256)
e031e1e2mo agomessage_footerNormalize decision support records at construction boundaries (#1255)
3cbd4f62mo agomessage_footerFix pass_priority(until=end_of_turn) timeout in golden replay (#1254)
84da57c2mo agomessage_footerExtract shared .surface-card CSS primitive (#1251)
5bcc34d2mo agomessage_footerUse plain dataclasses for decision support records (#1240)
fb91a242mo agomessage_footerMigrate misc export leaf records to dataclasses (#1244)
762f9202mo agomessage_footerMigrate Player/PilotPlayer from TypedDict to dataclass (#1245)
0a3d58e2mo agomessage_footerPlan TypedDict-to-dataclass migration (#1239)
deff5bb2mo agomessage_footerAdd deep analysis for game_20260308_225957_g2 (#1237)
ca3bbfe3mo agomessage_footerFail fast on dropped callback auto-responses (#1236)
73acff73mo agomessage_footerFail fast on remaining silent sendPlayer* failures (#1227)
a1d317d3mo agomessage_footerFix observer watch readiness race (#1228)
7e6ca163mo agomessage_footerFix lazy card loading: persist expansion metadata, support deck importers (#1226)
b3532df3mo agomessage_footerLazy per-card loading: skip bulk scan in observer and bridge (#1224)
ab3b6aa3mo agomessage_footerRe-throw delivery failures from getActionChoices (#1225)
2aef20b3mo agomessage_footerDeep analysis: game_20260308_234838_g1 (clean, no bugs) (#1223)
b3c15413mo agomessage_footerFail fast on silent sendPlayer* failures to prevent golden test hangs (#1214)
3ec1d453mo agomessage_footerFix blunder_eval --limit/--game crash and post-merge lint break (#1217)
bd3754c3mo agomessage_footerAdd CI watcher and fix orchestrator ARG001 lint (#1216)
0bb53843mo agomessage_footerTighten mypy: fix 38 suppressed errors, remove 3 module overrides (#1207)
6dad1253mo agomessage_footerCap JVM heap for all local XMage processes (#1215)
82460583mo agomessage_footerAdd metadata fields to shared MergedLlmEvent interface (#1213)
f428eb03mo agomessage_footerFix solve-issue skill: reference correct script name (#1210)
ec8f5c53mo agomessage_footerDeduplicate game log helpers; remove puppeteer dead code (#1212)
2d8fbd83mo agomessage_footerExtract shared CSS and replace hardcoded colors with variables (#1211)
4affb1e3mo agomessage_footerFile linting improvement issues (#1203)
6dacf8b3mo agomessage_footerPrerender game viewer: inline JSON + static action log (#1200)
b11e8b63mo agomessage_footerFix golden timeout: multi-thread bridge HTTP server (#1198)
52bace73mo agomessage_footerBump JavaFX from 11.0.2 to 21.0.10 (#1193)
a77dc3d3mo agomessage_footerFix TURN_MSG_PATTERN truncating parenthesized player names (#1196)
723d99a3mo agomessage_footerFix CI mcp-tools.json verification failure (#1192)
d0848663mo agomessage_footerFix HumanPlayer response window race (#1184)
cb5d25e3mo agomessage_footerUse instanceof pattern matching in bridge (#1183)
584ba923mo agomessage_footerFix nondeterministic MCP tools JSON ordering (#1185)
ce8a8c03mo agomessage_footerMigrate JAXB from javax.xml.bind to jakarta.xml.bind (#1177)
7b86c5b3mo agomessage_footerFix blunder eval mana-plan rendering (#1180)
909839d3mo agomessage_footer