AgentDeck is a research platform for studying AI behavior through game scenarios. Run controlled experiments with LLMs, collect comprehensive behavioral data, and replay matches for analysis. 🚧 Preview release - feedback welcome.
Stars
3
Forks
0
Watchers
3
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
68
commits
8
commits
Update all spec metadata to reflect Phase 6-8 completion
77134efView on GitHubFix CI failures: Python 3.9 compat, black formatting, bind_game bug
dcef6e9View on GitHubResolve merge conflicts: keep Phase 7/8 roadmap + CP2 reflection key
93e11c1View on GitHubResolve TL5/FP1/RE2/RE8 drift and harden Phase 8 tests
d3e5f08View on GitHubPhase 8: Add tests for Phase 7 invariants + fix 2 bugs
3e8e1cfView on GitHubRelax/drop unnecessary spec invariants based on Codex evaluation
5d0fa50View on GitHubUpdate ACTION-ITEMS.md: mark deferred items for research/spectator specs
8340db0View on GitHubFix SPEC-PRICING V0 critical drift: validate root type before .items()
1bff482View on GitHubFix SPEC-OBSERVABILITY (verified) and SPEC-PARALLEL (PO1 logging)
c5c6a07View on GitHubFix SPEC-MATCH-RUNTIME drift: pass game to handle_parse_failure
43a575aView on GitHub