v0.10.1

0.10.1 (2026-03-17)

Bug Fixes

• update tests for set-based Broadcaster connections and deterministic storm test (a6cf2ab)

Performance Improvements

• use set for Broadcaster connections — O(1) disconnect (#277) (499351e)
• use set for Broadcaster connections — O(1) disconnect (#277) (b1df69b)

v0.10.0

0.10.0 (2026-03-16)

Features

• harden server robustness with safe LLM response handling and security guards (05d90e2)

Bug Fixes

• add missing event_window_ticks to Settings (#269) (7a5a306)
• address CodeQL path injection warnings in SPA fallback and finetuning (6b8d6d9)
• harden agent loop error handling and tool dispatch (b887949)
• resolve CodeQL py/path-injection alerts with normpath/realpath+startswith pattern (b10f047)
• restore _reset_lock and update SPA guard test after rebase (6ea8bd1)
• restructure path guards so CodeQL recognizes startswith as barrier (c72b70d)

Performance Improvements

• use spatial index in _stone_proximity_concentration() (#270) (e4cef8b)

v0.9.0

0.9.0 (2026-03-12)

Features

• add automatic request_confirm for hazardous moves (261c58d)
• add automatic request_confirm for hazardous moves (e68c111)
• add Mistral Agents API as switchable backend (523b4ba)
• agents-api-threads: add persistent conversation threads with config toggle (e23ff80)
• agents-api-threads: persistent conversation threads + training logger (af348e7)
• agents-api: add Agents API reasoners for rover, drone, station (cee065c)
• agents-api: add loop variants and AGENT_MAP registration (506324b)
• config: add agent_backend toggle and tighten SDK constraint (500dc60)
• goal-confidence: add goal confidence tracking + UI bars (9973dc4)
• goal-confidence: add goal confidence tracking with UI bars (a34a57d)
• human-in-the-loop: add request_confirm rover tool with UI confirmation modal (3edc833)
• human-in-the-loop: add request_confirm rover tool with UI modal and confirmation flow (465136e)
• peer-messaging: add rover-to-rover direct messaging with notify_peer tool (52e430b)
• peer-messaging: rover-to-rover direct messaging with notify_peer tool (34d01a7)
• power-allocation: add allocate_power station tool with budget events and UI (583ca29)
• power-allocation: add allocate_power station tool with budget events and UI (6ccb6f5)
• presets: add multi-scenario simulation presets (3f3853e)
• presets: add multi-scenario simulation presets (df4eca9)
• replay: add simulation replay UI with training session playback (71bda7a)
• replay: add simulation replay UI with training session playback (2f78b49)

Bug Fixes

• ci: resolve eslint errors and flaky storm test (b827ea9)
• ci: resolve eslint errors and flaky storm test (890f656)
• resolve 8 simulation engine bugs (947f9a2)
• resolve 8 simulation engine bugs (tick inflation, tool whitelist, path checking, ice ratio, drone relay, station memory, geyser damage, storm multiplier) (8988d60)
• ui: remove unused requestId param in handleConfirmTimeout (9f06c63)

v0.8.0

0.8.0 (2026-03-02)

Features

• station: add recall command tool and host routing (#244) (8ee1819)

Bug Fixes

• ci: format station tests with ruff (#245) (2bef539)
• deps: patch vulnerable glob transitive dependency (#242) (5359497)

v0.7.1

0.7.1 (2026-03-02)

Bug Fixes

• ci: start/stop SurrealDB in pytest session hooks (#240) (3585f15)

v0.7.0

0.7.0 (2026-03-01)

Features

• add 5 new Mistral-model rovers (large, medium, codestral, ministral, magistral) (08e09b6)
• add ice/water/gas resource chain, Hauler agent, and base upgrades (b643129)
• ice/water/gas resource chain, Hauler agent, and base upgrades (e26ce4e)
• implement ice, water, gas plants and hauler agent (9a102cc)
• massive resource economy and hauler update (#235) (79360b4)

Bug Fixes

• correct hauler unload test assertion to match actual error message (25b8512)
• fix lint errors, add hauler tests, and format code (9337035)
• refine ice/gas/hauler mechanics and align tests with new schema (ce71d54)
• remove world.reset() calls in test_hauler (method does not exist) (02aa941)
• route hauler load_cargo/unload_cargo actions and remove stale imports (bdd19ec)
• stabilize resource economy — deduplicate constants, fix action dispatch, align tests (6ea0253)

v0.6.0

0.6.0 (2026-03-01)

Features

• add abandoned buildings and vehicles with obstacle, interaction, and passive effects (#124) (b243f25)
• auto-deliver rover inventory at station and show deposit events (#232) (210d353)
• enable narration by default and fix UI state sync (004d807)
• enable narration by default and fix UI state sync (ed7db33)
• Mars dust storm system — dynamic weather with battery drain and move interference (#196) (87dfa77)
• re-enable rover-2 on HuggingFace Qwen3-32B for multi-rover cooperation (27efa8c)
• re-enable rover-2 on HuggingFace Qwen3-32B for multi-rover cooperation (04ee7b4)
• simulation timer, session duration tracking, and control button verification (e4cd455)
• simulation timer, session duration tracking, and control button verification (ed3d05a)
• tune agent intervals, fix battery display, enhance map legend (7c779e9)
• tune agent intervals, fix battery display, enhance map legend (6dfc835)
• ui: enhance landing page visuals — mission loop icons, agent card animations, Mars globe atmosphere (878f909)
• ui: enhance landing page visuals — mission loop icons, agent cards, Mars globe (da451c9)

Bug Fixes

• increase agent intervals to avoid 429 rate limits (74109ce)
• increase agent intervals to avoid 429 rate limits (4s/3.5s) (cf0baa1)
• increase agent log pane height by 50% (1ea5e13)
• increase agent log pane height by 50% (30fa052)
• ui: tone down HowItWorks timeline animation line (552e940)
• ui: tone down HowItWorks timeline line — reduce opacity, thickness, and animation speed (f3952da)

Performance Improvements

• double agent action speed (c518714)
• double agent action speed (6s→3s rover, 5s→2.5s drone) (8acd3d5)

v0.5.0

0.5.0 (2026-03-01)

Features

• add environmental hazards — ice mountains and air geysers (#220) (f36cbdd)
• add SurrealDB training data logging for agent decision replay and fine-tuning (#212) (98b2d75)
• ui: add obstacle visualization to WorldMap SVG (#221) (36c9dfa)

Bug Fixes

• ci: resolve 3 CI failures blocking all branches (#216) (eddcddd)
• correct SyntaxError in training_logger, improve logging, add tests (#218) (8adfe61)
• resolve PR #145 runtime crashes in memory summarization (b7642d1)
• resolve PR #145 runtime crashes in memory summarization (#210) (94e5914)
• resolve production site issues — PORT, SPA routing, WS initial state (#219) (a64700f)
• tests: add assertions to tests that verified nothing (#142) (8323908)
• ui: guard AgentDetailModal against null properties (#89) (#217) (2d73524)

v0.4.0

0.4.0 (2026-03-01)

Features

• add HuggingFace Inference API as alternative LLM provider (#182) (937e572)
• add landing page with i18n, routing, and Three.js Mars globe (#206) (e174c49)

Bug Fixes

• address remaining bugs from stale PRs (#110, #152, #81, #78) (#211) (755e227)
• fix remaining charge_rover references to charge_agent (#68) (8afa9f3)
• fix remaining charge_rover references to charge_agent (#68) (a4bbebe)
• narrator: use async streaming to prevent event loop blocking (#203) (cbe5b7b)
• ui: preserve event log and simulation state on WebSocket reconnect (#72) (#209) (14e8f0a), closes #87
• use Settings.drone_turn_interval_seconds instead of hardcoded 2.0 (#98) (#208) (51f29ea)

v0.3.0

0.3.0 (2026-03-01)

Features

• add agent memory & learning system (Feature F) (4b32a9d)
• add agent reasoning transparency panel (Feature C) (64f333b)
• add live inter-agent communication (Feature A) (#150) (cf0ba69)
• add Station Reactive Intelligence (Feature E) (638fcbc)
• add world-data fine-tuning pipeline for continuous model improvement (80cd54c)
• add world-data fine-tuning pipeline for continuous model improvement (707d84c)
• Agent Memory & Learning (Feature F) (c7a1d06)
• Agent Reasoning Transparency Panel (Feature C) (24145b9)
• Station Reactive Intelligence (Feature E) — periodic LLM evaluation & coordination (1363a90)
• ui: add communication visualization on map (#126) (72bfd50)
• voice: add /api/voice-command endpoint with Voxtral transcription and LLM command parsing (#191) (ebfe37f)

Bug Fixes

• add thread-safe guards to Broadcaster disconnect/send (#100) (09e106a)
• add thread-safe guards to Broadcaster disconnect/send (#100) (e45c809)
• ci: apply ruff format to 7 unformatted files to fix server-lint (#185) (aa3f1a3)
• config: add pydantic Field validation for ports and intervals (f5d4502)
• config: add pydantic Field validation for ports and intervals (4d456f1), closes #119
• docker: add PYTHONUNBUFFERED and PYTHONDONTWRITEBYTECODE env vars (#180) (0936d2c)
• docker: add PYTHONUNBUFFERED and PYTHONDONTWRITEBYTECODE env vars (#180) (45567f7)
• ensure WebSocket disconnect cleanup via finally block (#101) (a195262)
• ensure WebSocket disconnect cleanup via finally block (#101) (09d31da)
• guard _random_free_pos against infinite loop (#118) (d663924)
• guard _random_free_pos against infinite loop (#118) (ccc2af4)
• harden Dockerfile with non-root user, pinned uv, and HEALTHCHECK (#117) (520ba45)
• harden Dockerfile with non-root user, pinned uv, and HEALTHCHECK (#117) (c846bde)
• remove legacy GRID_W/GRID_H boundary checks from rover _build_context() (085c7b2), closes #172
• remove legacy GRID_W/GRID_H boundary checks from rover context (#172) (0bfffe1)
• replace f-string logging and print() with lazy logger (#134, #135) (2b2be67)
• replace f-string logging and print() with lazy logger (#134, #135) (36307ac)
• tests: add tearDown to 5 test classes to prevent state leakage (4a3fa86)
• tests: add tearDown to 5 test classes to prevent state leakage (#174) (ce07ce7)
• tests: replace blocking sleep with asyncio.sleep, add file leak guard (#162) (8c2e1c3)
• tests: replace blocking sleep with asyncio.sleep, add file leak guard (#162) (86cbda3)
• ui: add cleanup for audio, toast timers, and animation timers on unmount (#94) (7ff9fd6)
• ui: add cleanup for audio, toast timers, and animation timers on unmount (#94) (92f61d0)
• ui: add error handling for WS JSON.parse and fetch calls (#95) (166525a)
• ui: add error handling for WS JSON.parse and fetch calls (#95) (93ba50b)
• ui: add error handling to SimulationPage fetch calls (5325720)
• ui: add error handling to SimulationPage fetch calls (1ff2d34), closes #121
• ui: add exponential backoff to WebSocket reconnect (68b8d73)
• ui: add exponential backoff to WebSocket reconnect (bfd93ed), closes #122
• ui: add flex-wrap and mobile breakpoint to MissionBar (#163) (7fe89e7)
• ui: add flex-wrap and mobile breakpoint to MissionBar (#163) (268d721)
• ui: add missing accessibility attributes across 5 components (c7e25e7)
• ui: add missing accessibility attributes across 5 components (01d6111), closes #140
• ui: batch accessibility, perf, and reactivity fixes (#164) (bf3cb8f)
• **ui:...