Adversarial 1v1 spotlight: ladder family + rating + Elo wiring

cargo tools: enter_transport + unload, 1:1 congruence

Unified Battle Viewer in app.py + run/model playback identity

guard tool: 1:1 congruence with engine Command.guard

Wire goal tracker into scoring + leaderboard

Playback: capture model reasoning + per-turn goal tracker + viewer

Scenario-controlled tool allow/deny + default core set

S7 bench: set_stance + patrol tools (schema 1:1, 17==17)

S7 bench: surrender tool + loss outcome (tool schema 1:1, 15==15)

Step 4 (bench): interrupt-driven loop + playback/trace capture

Step 2: concurrent scenario execution (--concurrency N)

Pipeline step 7: per-episode playback persistence

Bench: surface S9 spatial tensor in render_state (multimodal reach)

Pairwise adversarial eval + Elo (the user's 'pairwise conditions')

Generalization-gap metric: held-out split in run_eval + leaderboard

Catalog C13: harvest-economy packs — closes #14 (user economy families)

S1 bench wiring: resources/capacity + economy_value win predicates

Research-grounded 200-level scenario catalog (12 categories) + integ

Scenario brainstorm + 2 runnable economy packs

Single source of truth: port rush-hour + 3 strategy scenarios to packs/

adapter: use engine map_info for true map dims (S9), synthesis fallback

Custom-map no-enemy scenario pack + robustness regression

#13: robustness regression suite (fail-safe + determinism)

Building & Planning scenario family (user-specified first set)

#12: real custom-map terrain via dynamic .oramap registry

#6 leaderboard: data layer + run_eval publish + Gradio tab

Bench: economy scenario pack + full-loop integ test + starting_cash constraint

tests: reset reveals sight (engine fix); robust in-bounds move targets

Bench: consume S9 economy obs + economy win-conditions + full toolset

Add scoring + P/R/A diagnostics + run_eval CLI

Add provider-agnostic model agent (vLLM/OpenRouter/Bedrock)

Add Rust-backed eval stack: scenario packs, adapter, spine, integration tests

Add HF identity verification and anonymous submission support

Add server-side game aggregation with minimum 5-game threshold

Add minimum game count, difficulty multiplier, and updated scoring formula

Security hardening: XSS prevention, input validation, rate limiting

Update docs: CLI submission, agent identity, replay downloads, API endpoints

Add agent URL hyperlinks, replay downloads, and submit_with_replay endpoint

Merge pull request #3 from yxc20089/feature/upload-form-5tiers

Fix CI: remove openra_env dependency from get_agent_fn

Fix CI: add openra-rl-util to requirements.txt

Revert "Fix CI: import scoring from evaluate_runner instead of openra_rl_util"

Fix CI: import scoring from evaluate_runner instead of openra_rl_util

Add upload form, API endpoint, 5 difficulty tiers, real game data

Move Try experience to OpenRA-RL Space, remove Try tab from Bench

Rename Evaluate tab to Try, stream LLM agent gameplay

Add in-browser evaluation via Evaluate tab

Connect evaluation harness to HF-hosted OpenRA-RL environment

Fix HF Space build: remove unused deps from requirements.txt

Move HF Space sync to openra-rl org