Changes behind the numbers
Every change that reached a project's default branch. Greyed rows were not counted, with the reason. A project from an older card may show dates and line counts only.
Fort Gym
LIVEUNCHECKED- Merge reviewed Year-Two autonomous play releasea3e4963110817 lines
- Merge pull request #136 from lemoz/codex/campaign-map-inspection#1362989 lines
- Merge pull request #135 from lemoz/codex/campaign-observation-context#135399 lines
- Merge pull request #134 from lemoz/codex/campaign-dialog-feedback#134596 lines
- Merge pull request #132 from lemoz/codex/native-output-pause-acceptance#1323083 lines
- Merge pull request #131 from lemoz/codex/campaign-capacity-preflight#131380 lines
- Merge pull request #130 from lemoz/codex/native-campaign-runtime#13018549 lines
- Merge pull request #129 from lemoz/codex/campaign-agent-foundation#1292349 lines
- Merge pull request #128 from lemoz/codex/long-v2-results#1281493 lines
- Merge pull request #127 from lemoz/codex/native-drink-inventory#127402 lines
- Merge pull request #126 from lemoz/codex/campaign-website-delivery#12612297 lines
- Merge pull request #124 from lemoz/codex/governed-contract-recovery#124551 lines
- Merge pull request #123 from lemoz/codex/meeting-interrupt-recovery#123116 lines
- Fix canonical DF save resolution after seed reset (#122)#12241 lines
- feat: add Fort-Eval Easy P1 frontier comparison (#121)#1212659 lines
- Add Fort Labs social previews (#120)#1201109 lines
- Add evidence-backed Fort Labs findings (#119)#119416 lines
- Redesign the Fort Labs replay workspace (#118)#118618 lines
- Add protocol-scoped Fort Labs results (#117)#1171132 lines
- Merge pull request #116 from lemoz/codex/fort-labs-site-v1#1162268 lines
- Merge pull request #115 from lemoz/codex/fort-eval-v0.1#1152868 lines
- Merge pull request #114 from lemoz/codex/governed-review-repair#114189 lines
- Merge pull request #113 from lemoz/codex/labor-eligible-citizens#113831 lines
- Merge pull request #112 from lemoz/codex/attempt27-findings#112124 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #111 from lemoz/codex/g7-citizen-ledger#111108 lines
- Merge pull request #110 from lemoz/codex/governed-objective-completion#110778 lines
- Merge pull request #109 from lemoz/codex/farm-contained-evidence#109311 lines
- Merge pull request #108 from lemoz/codex/transition-repause-attestation#10896 lines
- Merge pull request #107 from lemoz/codex/atomic-sample-json#10764 lines
- Merge pull request #106 from lemoz/codex/attempt27-modal-interrupt#1063246 lines
- Merge pull request #105 from lemoz/codex/attempt27-completion-feedback#105146 lines
- Merge pull request #104 from lemoz/codex/attempt27-phase-transition#10414 lines
- Merge pull request #103 from lemoz/codex/attempt26-truth-fixes#103667 lines
- Merge pull request #102 from lemoz/codex/attempt26-launch-proof#10240 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #101 from lemoz/codex/attempt25-observation-fixes#101269 lines
- Merge pull request #100 from lemoz/codex/post-pause-final-tick#10067 lines
- Merge pull request #99 from lemoz/codex/fail-closed-tick-controller#99942 lines
- Merge pull request #98 from lemoz/codex/job-target-connectivity#98120 lines
- Merge pull request #97 from lemoz/codex/full-workshop-footprint-feedback#97241 lines
- Merge pull request #96 from lemoz/codex/attempt23-live-proof#9660 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #95 from lemoz/codex/attempt22-stores-exit#95701 lines
- Merge pull request #94 from lemoz/codex/attempt22-g7-facts#94412 lines
- Merge pull request #93 from lemoz/codex/attempt22-feedback-truth#93153 lines
- Merge pull request #92 from lemoz/codex/attempt21-agent-observation#92338 lines
- Merge pull request #91 from lemoz/codex/attempt20-fort-map-anchor#91403 lines
- Merge pull request #90 from lemoz/codex/subterranean-access#905780 lines
- Merge pull request #89 from lemoz/codex/attempt20-action-truth#891721 lines
- Merge pull request #88 from lemoz/codex/attempt19-truthful-perception#88259 lines
- Merge pull request #87 from lemoz/codex/governed-parallel-review#87495 lines
- Merge pull request #86 from lemoz/codex/focus-review-corrections#86371 lines
- Merge pull request #85 from lemoz/codex/review-evidence-correction#85276 lines
- Merge pull request #84 from lemoz/codex/glm52-json-headroom#84105 lines
- Merge pull request #83 from lemoz/codex/glm52-json-transport#83165 lines
- Merge pull request #82 from lemoz/codex/glm52-auto-tool-choice#82145 lines
- Merge pull request #81 from lemoz/codex/occupied-map-farm-room-contract#81330 lines
- Expose stable workshop sites and frozen-liquid truth (#80)#80991 lines
- agent: clarify objective correction decisions (#79)#79196 lines
- agent: give GLM-5V review output headroom (#78)#78132 lines
- agent: aggregate cross-stage correction errors (#77)#77211 lines
- agent: preserve governed correction context (#76)#76473 lines
- Harden governed review resilience (#75)#75206 lines
- Require factual governed plan reviews (#74)#742040 lines
- Ground governed BUILD targets in observed map (#73)#7323 lines
- Expose factual governed action history (#72)#72801 lines
- docs: record G7 attempt 5 launch (#71)#7157 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- dfhack: harden executable work and dialog proof (#70)#701484 lines
- dfhack: harden workshop and run finalization (#69)#691294 lines
- dfhack: repair G7 evidence and seed loading (#68)#68394 lines
- dfhack: harden G7 run integrity and evidence (#67)#674292 lines
- WDSLL: ratify G7 at the 2026-07-09 window; pre-declare attempt-1 scalar bar (#6…#6624 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- HOLD FOR OPERATOR WINDOW (G7 primitive): LABOR — agent-controlled labor assignm…#65920 lines
- HOLD FOR OPERATOR WINDOW (G7 primitive): FARM — agent-controlled crop selection…#64955 lines
- Fix brew ORDER: BrewDrink does not exist on 0.47.05 — use CustomReaction (#63)#6395 lines
- HOLD FOR OPERATOR WINDOW (G7 primitive): Still workshop + brew ORDER (#56)#56432 lines
- HOLD FOR OPERATOR WINDOW (G7 primitive): farm-plot BUILD kind (#54)#54573 lines
- HOLD FOR OPERATOR WINDOW (G7 primitive): gather-plants DIG kind (#53)#53424 linesfolded into the change before it
- Segment public leaderboard by score_version + seed_save (#61)#61507 linesfolded into the change before it
- docs: FINDINGS.md — the lab's findings, from the corrections log (#62)#62363 linesfolded into the change before it
- docs: G6 escalation log — first rooms on the unseen map (#60)#6039 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- docs: record operator escalation choice — hybrid GPT-5.5-vision + corrected GLM…e2fe2c16 linesfolded into the change before it
- Merge pull request #59 from lemoz/fix/usable-stocks-pipeline#5981 lines
- Merge pull request #58 from lemoz/fix/usable-stocks#58112 lines
- Merge pull request #57 from lemoz/feat/landing-page#57462 linesfolded into the change before it
- Merge pull request #55 from lemoz/fix/g6-attempt1-corrections#55193 lines
- Merge pull request #52 from lemoz/docs/g6-protocol#5220 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #51 from lemoz/feat/per-run-seed-selection#51130 lines
- Merge pull request #47 from lemoz/docs/g4-repeatability-protocol#4746 linesfolded into the change before it
- Merge pull request #43 from lemoz/docs/g4-first-pass#4323 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #48 from lemoz/docs/score-v3-and-g7-proposal#48109 linesfolded into the change before it
- Merge pull request #50 from lemoz/feat/score-v3#501635 lines
- Merge pull request #49 from lemoz/fix/ci-agent-extra#492 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #46 from lemoz/fix/fort-following-snapshot-crew-survey#46164 lines
- Merge pull request #45 from lemoz/fix/replay-fort-minimap#45247 linesfolded into the change before it
- Merge pull request #44 from lemoz/fix/fort-following-snapshot#44123 lines
- Merge pull request #42 from lemoz/docs/g4-run3-log#4221 linesfolded into the change before it
- Merge pull request #41 from lemoz/feat/unsuspend-action#41334 lines
- Merge pull request #39 from lemoz/docs/g4-vision-series-log#3943 linesfolded into the change before it
- Merge pull request #40 from lemoz/fix/furniture-tile-errors#4040 lines
- Merge pull request #36 from lemoz/fix/serialize-dfhack-jobs#3618 linesfolded into the change before it
- Merge pull request #38 from lemoz/fix/queued-construction-visibility#38109 lines
- Merge pull request #35 from lemoz/fix/decouple-legacy-plan#35434 lines
- Merge pull request #34 from lemoz/feature/replay-bigger-map#3486 lines
- docs: record vision experiment — rings are a modality problemfd1cda916 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge pull request #33 from lemoz/fix/kimi-token-budget#3331 lines
- Merge pull request #32 from lemoz/fix/text-json-action-fallback#3265 lines
- Merge pull request #31 from lemoz/fix/reasoning-mandatory#3153 lines
- Merge pull request #30 from lemoz/feature/replay-graphical-glyphs#30264 lines
- Merge pull request #29 from lemoz/feature/vision-arms#2979 lines
- Merge pull request #28 from lemoz/fix/vision-provider-quirks#28107 lines
- Merge pull request #27 from lemoz/feature/vision-minimap#27213 lines
- Merge pull request #22 from lemoz/codex/real-map-tile-replay#22464 linesfolded into the change before it
- Merge pull request #26 from lemoz/merge/4-plan-agnostic#261535 linesfolded into the change before it
- Merge pull request #25 from lemoz/merge/3-gates-g2-g3#251039 linesfolded into the change before it
- Merge pull request #24 from lemoz/merge/2-governed-llm#242019 linesfolded into the change before it
- Merge pull request #23 from lemoz/merge/1-replay-evidence#2323208 lines
- Render saved run replay from trace0dc7e6b577 lines
- Recover stale public screenshot clients370ac57101 lines
- Score visible fortress complexity833c834679 lines
- Prove live production progress414ec97479 lines
- Avoid stale cached live UI (#17)#1729 lines
- Improve live run inspection UI (#16)#1689 lines
- [codex] Add live utility progress metrics (#15)#15524 lines
- Add live mining completion metrics5f7439b454 lines
- Add live work progress metrics33712de588 lines
- Add live agent suite scorecard (#12)#12671 lines
- Add visible live agent comparison (#11)#11544 lines
- Add live demo rehearsal commandf05a2d2325 lines
- Add live DFHack smoke pathbd6c4e8489 lines
- Add Claude Sonnet mock run example (#8)#8236 lines
- Add mock scenario pack assertions (#7)#7332 lines
- Add interview demo script (#6)#6156 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Use Node 24 GitHub Actions (#5)#54 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Add GitHub Actions CI (#4)#433 lines
- Add leaderboard screenshot to README (#3)#32 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Add mock quickstart path (#2)#2156 lines
- Merge pull request #1 from lemoz/readme-hero#113 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge WO-004: manual resolutiona6b792a425 linesfolded into the change before it
- Merge WO-003: manual resolutionb3f4b71519 lines
- Merge WO-002: manual resolutioneb26cae207 linesfolded into the change before it
- Merge WO-001: manual resolution9997ba235 lines
- agent: emphasize horizontal navigation to mountain before z-levels19f390338 lines
- agent: clarify screen borders vs stone, add z-level navigatione872bb245 lines
- agent: explain 3-column screen layout to agent876067188 lines
- env: add screen diff feedback to observations5dce0c878 lines
- run: enable screen capture for -research agents870f4963 linesfolded into the change before it
- api: add anthropic-research to ModelType schema6671ad92 linesfolded into the change before it
- agent: add anthropic-research agent with web search capabilityada5727342 linesfolded into the change before it
- feat(agent): add agent-controlled time with advance_ticksb6353e164 lines
- docs: switch subagent examples to gpt-5.2-codex0238a8310 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- docs: add xhigh reasoning subagent example99c81155 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- docs: align subagent commands with Codex CLI docsa83c23e23 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- docs: add tmux + worktree subagent runsd40d47131 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- docs: add CEO + subagent brief workflow5889b5751 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- docs: clarify VM user-folder workflow043058944 linesnot counted: 5 or fewer lines, or docs only, and no one else reviewed it
- Merge branch 'feature/prod-hardening-persistence-leaderboard'a45e5ed1676 lines
- docs: refresh VM seed workflow6133fb149 linesfolded into the change before it
- api: unify artifacts directoryb5483449 linesfolded into the change before it
- dfhack: reset runs from seed savesb40dfb5212 lines
- dfhack: handle seed reset permissionsfd7d71732 linesfolded into the change before it
- docs: add DFHack seed reset workflow2fa0f3726 linesfolded into the change before it
- dfhack: reset current save from seed36847ae150 lines
- dfhack: translate embark hotkeye37e49a38 lines
- api,web: add admin basic auth and keyboard control4d6f09f187 lines
- docs: update changelog for v0.4.03b1ea8318 linesfolded into the change before it
- infra: add vm-test/vm-deploy make targetsa4782e031 linesfolded into the change before it
- docs: document drift-free VM deployment workflowdbf276449 linesfolded into the change before it
- infra: add DFHack recovery scripts and build roled430c361130 linesfolded into the change before it
- feat(eval,env): add pause feedback and trace analyzer3a0f037745 lines
- fix(api,web): auto-create share tokens and fix stats displayc49e19c75 lines
- fix(dfhack): correct population/food/drink/wealth state readingdc0d4641364 lines
- feat(web): add Live Game View to admin panel with pcface CP437 renderingdb1b06b431 lines
- feat(agent): improve keystroke mode with action history and pause-on-endc3ea615104 lines
- feat(agent): add keystroke control mode for Claude playing DF like a humanaf65a18701 lines
- feat(web): add Live Game View with CopyScreen API6b486af210 lines
- dfhack: fix tick advancement and ORDER actions for v0.47.056fe1773563 lines