@opengamehq/cli 0.9.7 → 0.9.9
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/AGENTS.md +37 -17
- package/CHANGELOG.md +18 -0
- package/README.md +7 -6
- package/dist/cli.mjs +144 -144
- package/dist/opengame-global.js +3 -3
- package/dist/preflight-worker.mjs +14 -14
- package/dist/runtime-builder.mjs +1 -1
- package/dist/skills/manifest.json +1406 -142
- package/dist/skills/opengame-accessibility/SKILL.md +48 -32
- package/dist/skills/opengame-all-in-one/SKILL.md +195 -0
- package/dist/skills/opengame-all-in-one/references/opengame-accessibility/README.md +102 -0
- package/dist/skills/opengame-all-in-one/references/opengame-api/README.md +107 -0
- package/dist/skills/opengame-all-in-one/references/opengame-assets/README.md +168 -0
- package/dist/skills/opengame-all-in-one/references/opengame-assets/references/pipeline.md +164 -0
- package/dist/skills/opengame-all-in-one/references/opengame-audio/README.md +238 -0
- package/dist/skills/opengame-all-in-one/references/opengame-audio/helpers/audio-chain.js +325 -0
- package/dist/skills/opengame-all-in-one/references/opengame-audio/references/mix-notes.md +125 -0
- package/dist/skills/opengame-all-in-one/references/opengame-audio/references/qa-without-ears.md +177 -0
- package/dist/skills/opengame-all-in-one/references/opengame-capture/README.md +128 -0
- package/dist/skills/opengame-all-in-one/references/opengame-capture/references/pipeline.md +114 -0
- package/dist/skills/opengame-all-in-one/references/opengame-capture/references/threejs.md +161 -0
- package/dist/skills/opengame-all-in-one/references/opengame-characters/README.md +180 -0
- package/dist/skills/opengame-all-in-one/references/opengame-characters/references/first-person-body.md +108 -0
- package/dist/skills/opengame-all-in-one/references/opengame-characters/references/integration.md +144 -0
- package/dist/skills/opengame-all-in-one/references/opengame-characters/references/player-og.md +160 -0
- package/dist/skills/opengame-all-in-one/references/opengame-cli/README.md +189 -0
- package/dist/skills/opengame-all-in-one/references/opengame-deploy/README.md +223 -0
- package/dist/skills/opengame-all-in-one/references/opengame-deploy/references/json-events.md +63 -0
- package/dist/skills/opengame-all-in-one/references/opengame-deploy/references/manifest.md +89 -0
- package/dist/skills/opengame-all-in-one/references/opengame-deploy/references/packaging.md +119 -0
- package/dist/skills/opengame-all-in-one/references/opengame-depth/README.md +154 -0
- package/dist/skills/opengame-all-in-one/references/opengame-enemies/README.md +155 -0
- package/dist/skills/opengame-all-in-one/references/opengame-enemies/references/pacing-and-bots.md +108 -0
- package/dist/skills/opengame-all-in-one/references/opengame-enemies/references/threejs.md +193 -0
- package/dist/skills/opengame-all-in-one/references/opengame-front-end/README.md +185 -0
- package/dist/skills/opengame-all-in-one/references/opengame-front-end/references/launch.md +50 -0
- package/dist/skills/opengame-all-in-one/references/opengame-front-end/references/menus.md +81 -0
- package/dist/skills/opengame-all-in-one/references/opengame-game-feel/README.md +181 -0
- package/dist/skills/opengame-all-in-one/references/opengame-game-feel/helpers/spring.js +146 -0
- package/dist/skills/opengame-all-in-one/references/opengame-game-feel/references/feel-numbers.md +142 -0
- package/dist/skills/opengame-all-in-one/references/opengame-game-feel/references/killcam.md +88 -0
- package/dist/skills/opengame-all-in-one/references/opengame-jam/README.md +114 -0
- package/dist/skills/opengame-all-in-one/references/opengame-lighting/README.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-lighting/helpers/light-pool.js +185 -0
- package/dist/skills/opengame-all-in-one/references/opengame-loading/README.md +135 -0
- package/dist/skills/opengame-all-in-one/references/opengame-loading/helpers/loading-count.mjs +115 -0
- package/dist/skills/opengame-all-in-one/references/opengame-loading/references/threejs.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-look/README.md +181 -0
- package/dist/skills/opengame-all-in-one/references/opengame-look/references/matching-a-reference.md +73 -0
- package/dist/skills/opengame-all-in-one/references/opengame-look/references/painterly.md +69 -0
- package/dist/skills/opengame-all-in-one/references/opengame-look/references/post-and-lens.md +73 -0
- package/dist/skills/opengame-all-in-one/references/opengame-maps/README.md +74 -0
- package/dist/skills/opengame-all-in-one/references/opengame-maps/references/map-spec.md +77 -0
- package/dist/skills/opengame-all-in-one/references/opengame-mobile/README.md +86 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/README.md +183 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/briefs.md +293 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/day-one-questions.md +46 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/edge-cases.md +27 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/filmable.md +244 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/finishing.md +126 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/measure.md +282 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/montage-capture.md +151 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/montage-gates.md +84 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/montage-genres.md +198 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/montage-grammar.md +320 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/montage-ops.md +127 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/og-snipers-record.md +110 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/owner-eye.md +214 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/reference-study.md +293 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-make/references/song.md +64 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/README.md +193 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/briefs.md +293 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/day-one-questions.md +46 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/filmable.md +244 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/measure.md +282 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/montage-capture.md +151 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/montage-gates.md +84 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/montage-genres.md +198 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/montage-grammar.md +320 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/montage-ops.md +127 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/og-snipers-record.md +110 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/owner-eye.md +214 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/reference-study.md +293 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-plan/references/song.md +64 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/README.md +174 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/briefs.md +293 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/day-one-questions.md +46 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/filmable.md +244 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/finishing.md +126 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/measure.md +282 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/montage-capture.md +151 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/montage-gates.md +84 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/montage-genres.md +198 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/montage-grammar.md +320 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/montage-ops.md +127 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/og-snipers-record.md +110 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/owner-eye.md +214 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/reference-study.md +293 -0
- package/dist/skills/opengame-all-in-one/references/opengame-montage-review-and-improve/references/song.md +64 -0
- package/dist/skills/opengame-all-in-one/references/opengame-multiplayer/README.md +139 -0
- package/dist/skills/opengame-all-in-one/references/opengame-multiplayer/references/bot-lobbies.md +71 -0
- package/dist/skills/opengame-all-in-one/references/opengame-multiplayer/references/netcode.md +165 -0
- package/dist/skills/opengame-all-in-one/references/opengame-performance/README.md +177 -0
- package/dist/skills/opengame-all-in-one/references/opengame-performance/helpers/landing-pacer.mjs +86 -0
- package/dist/skills/opengame-all-in-one/references/opengame-performance/references/measuring.md +123 -0
- package/dist/skills/opengame-all-in-one/references/opengame-performance/references/webgpu.md +86 -0
- package/dist/skills/opengame-all-in-one/references/opengame-production/README.md +192 -0
- package/dist/skills/opengame-all-in-one/references/opengame-production/references/team.md +107 -0
- package/dist/skills/opengame-all-in-one/references/opengame-production/references/templates.md +153 -0
- package/dist/skills/opengame-all-in-one/references/opengame-quality-bar/README.md +170 -0
- package/dist/skills/opengame-all-in-one/references/opengame-quality-bar/references/cheap-tells.md +43 -0
- package/dist/skills/opengame-all-in-one/references/opengame-quality-bar/references/judges.md +95 -0
- package/dist/skills/opengame-all-in-one/references/opengame-quality-bar/references/review-prompt.md +32 -0
- package/dist/skills/opengame-all-in-one/references/opengame-rhythm/README.md +134 -0
- package/dist/skills/opengame-all-in-one/references/opengame-rhythm/references/charting.md +109 -0
- package/dist/skills/opengame-all-in-one/references/opengame-rhythm/references/timing.md +92 -0
- package/dist/skills/opengame-all-in-one/references/opengame-router/README.md +54 -0
- package/dist/skills/opengame-all-in-one/references/opengame-router/references/process.md +38 -0
- package/dist/skills/opengame-all-in-one/references/opengame-scene-audit/README.md +136 -0
- package/dist/skills/opengame-all-in-one/references/opengame-scene-audit/helpers/scene-audit.js +677 -0
- package/dist/skills/opengame-all-in-one/references/opengame-sdk/README.md +283 -0
- package/dist/skills/opengame-all-in-one/references/opengame-sdk/references/engines.md +93 -0
- package/dist/skills/opengame-all-in-one/references/opengame-sdk/references/frame.md +107 -0
- package/dist/skills/opengame-all-in-one/references/opengame-sdk/references/identity.md +111 -0
- package/dist/skills/opengame-all-in-one/references/opengame-sdk/references/leaderboards.md +115 -0
- package/dist/skills/opengame-all-in-one/references/opengame-sdk/references/radio.md +71 -0
- package/dist/skills/opengame-all-in-one/references/opengame-stability/README.md +112 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/README.md +160 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/archetypes.md +113 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/briefs.md +273 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/core.md +129 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/craft.md +299 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/edge-cases.md +27 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/filmable.md +244 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/finishing.md +126 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/gates.md +184 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/measure.md +282 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/music.md +148 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/og-clash-record.md +133 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/ops.md +154 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/owner-eye.md +214 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/review.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-make/references/title.md +92 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/README.md +172 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/archetypes.md +113 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/briefs.md +273 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/core.md +129 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/craft.md +299 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/edge-cases.md +27 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/filmable.md +244 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/finishing.md +126 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/gates.md +184 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/measure.md +282 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/music.md +148 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/og-clash-record.md +133 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/ops.md +154 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/owner-eye.md +214 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/reference-set.md +92 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/review.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-plan/references/title.md +92 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/README.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/archetypes.md +113 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/briefs.md +273 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/core.md +129 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/craft.md +299 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/edge-cases.md +27 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/filmable.md +244 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/finishing.md +126 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/gates.md +184 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/measure.md +282 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/music.md +148 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/og-clash-record.md +133 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/ops.md +154 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/owner-eye.md +214 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/review.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-trailer-review-and-improve/references/title.md +92 -0
- package/dist/skills/opengame-all-in-one/references/opengame-verification/README.md +197 -0
- package/dist/skills/opengame-all-in-one/references/opengame-verification/helpers/browser-evidence.mjs +74 -0
- package/dist/skills/opengame-all-in-one/references/opengame-weapons/README.md +179 -0
- package/dist/skills/opengame-all-in-one/references/opengame-weapons/references/handling-numbers.md +114 -0
- package/dist/skills/opengame-all-in-one/references/opengame-weapons/references/reloads-and-hands.md +96 -0
- package/dist/skills/opengame-all-in-one/references/opengame-weapons/references/sights-and-scopes.md +98 -0
- package/dist/skills/opengame-all-in-one/references/opengame-weapons/references/threejs.md +240 -0
- package/dist/skills/opengame-all-in-one/references/opengame-world-physics/README.md +141 -0
- package/dist/skills/opengame-all-in-one/references/opengame-world-physics/references/measured-failures.md +79 -0
- package/dist/skills/opengame-api/SKILL.md +32 -6
- package/dist/skills/opengame-assets/SKILL.md +79 -31
- package/dist/skills/opengame-assets/references/pipeline.md +67 -0
- package/dist/skills/opengame-audio/SKILL.md +129 -41
- package/dist/skills/opengame-audio/references/mix-notes.md +44 -1
- package/dist/skills/opengame-audio/references/qa-without-ears.md +177 -0
- package/dist/skills/opengame-capture/SKILL.md +44 -20
- package/dist/skills/opengame-capture/references/pipeline.md +33 -1
- package/dist/skills/opengame-capture/references/threejs.md +17 -1
- package/dist/skills/opengame-characters/SKILL.md +94 -39
- package/dist/skills/opengame-characters/references/first-person-body.md +108 -0
- package/dist/skills/opengame-characters/references/player-og.md +1 -1
- package/dist/skills/opengame-cli/SKILL.md +4 -4
- package/dist/skills/opengame-deploy/SKILL.md +35 -15
- package/dist/skills/opengame-deploy/references/json-events.md +6 -5
- package/dist/skills/opengame-deploy/references/packaging.md +28 -26
- package/dist/skills/opengame-depth/SKILL.md +61 -33
- package/dist/skills/opengame-enemies/SKILL.md +96 -40
- package/dist/skills/opengame-enemies/references/pacing-and-bots.md +108 -0
- package/dist/skills/opengame-enemies/references/threejs.md +59 -0
- package/dist/skills/opengame-front-end/SKILL.md +85 -39
- package/dist/skills/opengame-front-end/references/menus.md +42 -8
- package/dist/skills/opengame-game-feel/SKILL.md +95 -47
- package/dist/skills/opengame-game-feel/references/feel-numbers.md +59 -0
- package/dist/skills/opengame-game-feel/references/killcam.md +88 -0
- package/dist/skills/opengame-jam/SKILL.md +42 -27
- package/dist/skills/opengame-lighting/SKILL.md +85 -34
- package/dist/skills/opengame-loading/SKILL.md +78 -38
- package/dist/skills/opengame-loading/helpers/loading-count.mjs +115 -0
- package/dist/skills/opengame-loading/references/threejs.md +32 -7
- package/dist/skills/opengame-look/SKILL.md +122 -51
- package/dist/skills/opengame-look/references/matching-a-reference.md +73 -0
- package/dist/skills/opengame-look/references/painterly.md +69 -0
- package/dist/skills/opengame-look/references/post-and-lens.md +10 -3
- package/dist/skills/opengame-maps/SKILL.md +78 -0
- package/dist/skills/opengame-maps/references/map-spec.md +77 -0
- package/dist/skills/opengame-mobile/SKILL.md +18 -18
- package/dist/skills/opengame-montage-make/SKILL.md +187 -0
- package/dist/skills/opengame-montage-make/references/briefs.md +293 -0
- package/dist/skills/opengame-montage-make/references/day-one-questions.md +46 -0
- package/dist/skills/opengame-montage-make/references/edge-cases.md +27 -0
- package/dist/skills/opengame-montage-make/references/filmable.md +244 -0
- package/dist/skills/opengame-montage-make/references/finishing.md +126 -0
- package/dist/skills/opengame-montage-make/references/measure.md +282 -0
- package/dist/skills/opengame-montage-make/references/montage-capture.md +151 -0
- package/dist/skills/opengame-montage-make/references/montage-gates.md +84 -0
- package/dist/skills/opengame-montage-make/references/montage-genres.md +198 -0
- package/dist/skills/opengame-montage-make/references/montage-grammar.md +320 -0
- package/dist/skills/opengame-montage-make/references/montage-ops.md +127 -0
- package/dist/skills/opengame-montage-make/references/og-snipers-record.md +110 -0
- package/dist/skills/opengame-montage-make/references/owner-eye.md +214 -0
- package/dist/skills/opengame-montage-make/references/reference-study.md +293 -0
- package/dist/skills/opengame-montage-make/references/song.md +64 -0
- package/dist/skills/opengame-montage-plan/SKILL.md +197 -0
- package/dist/skills/opengame-montage-plan/references/briefs.md +293 -0
- package/dist/skills/opengame-montage-plan/references/day-one-questions.md +46 -0
- package/dist/skills/opengame-montage-plan/references/filmable.md +244 -0
- package/dist/skills/opengame-montage-plan/references/measure.md +282 -0
- package/dist/skills/opengame-montage-plan/references/montage-capture.md +151 -0
- package/dist/skills/opengame-montage-plan/references/montage-gates.md +84 -0
- package/dist/skills/opengame-montage-plan/references/montage-genres.md +198 -0
- package/dist/skills/opengame-montage-plan/references/montage-grammar.md +320 -0
- package/dist/skills/opengame-montage-plan/references/montage-ops.md +127 -0
- package/dist/skills/opengame-montage-plan/references/og-snipers-record.md +110 -0
- package/dist/skills/opengame-montage-plan/references/owner-eye.md +214 -0
- package/dist/skills/opengame-montage-plan/references/reference-study.md +293 -0
- package/dist/skills/opengame-montage-plan/references/song.md +64 -0
- package/dist/skills/opengame-montage-review-and-improve/SKILL.md +178 -0
- package/dist/skills/opengame-montage-review-and-improve/references/briefs.md +293 -0
- package/dist/skills/opengame-montage-review-and-improve/references/day-one-questions.md +46 -0
- package/dist/skills/opengame-montage-review-and-improve/references/filmable.md +244 -0
- package/dist/skills/opengame-montage-review-and-improve/references/finishing.md +126 -0
- package/dist/skills/opengame-montage-review-and-improve/references/measure.md +282 -0
- package/dist/skills/opengame-montage-review-and-improve/references/montage-capture.md +151 -0
- package/dist/skills/opengame-montage-review-and-improve/references/montage-gates.md +84 -0
- package/dist/skills/opengame-montage-review-and-improve/references/montage-genres.md +198 -0
- package/dist/skills/opengame-montage-review-and-improve/references/montage-grammar.md +320 -0
- package/dist/skills/opengame-montage-review-and-improve/references/montage-ops.md +127 -0
- package/dist/skills/opengame-montage-review-and-improve/references/og-snipers-record.md +110 -0
- package/dist/skills/opengame-montage-review-and-improve/references/owner-eye.md +214 -0
- package/dist/skills/opengame-montage-review-and-improve/references/reference-study.md +293 -0
- package/dist/skills/opengame-montage-review-and-improve/references/song.md +64 -0
- package/dist/skills/opengame-multiplayer/SKILL.md +143 -0
- package/dist/skills/opengame-multiplayer/references/bot-lobbies.md +71 -0
- package/dist/skills/opengame-multiplayer/references/netcode.md +165 -0
- package/dist/skills/opengame-performance/SKILL.md +106 -65
- package/dist/skills/opengame-performance/helpers/landing-pacer.mjs +86 -0
- package/dist/skills/opengame-performance/references/measuring.md +51 -0
- package/dist/skills/opengame-performance/references/webgpu.md +86 -0
- package/dist/skills/opengame-production/SKILL.md +119 -54
- package/dist/skills/opengame-production/references/team.md +107 -0
- package/dist/skills/opengame-production/references/templates.md +106 -16
- package/dist/skills/opengame-quality-bar/SKILL.md +94 -42
- package/dist/skills/opengame-quality-bar/references/cheap-tells.md +14 -7
- package/dist/skills/opengame-quality-bar/references/judges.md +95 -0
- package/dist/skills/opengame-quality-bar/references/review-prompt.md +10 -5
- package/dist/skills/opengame-rhythm/SKILL.md +138 -0
- package/dist/skills/opengame-rhythm/references/charting.md +109 -0
- package/dist/skills/opengame-rhythm/references/timing.md +92 -0
- package/dist/skills/opengame-router/SKILL.md +10 -2
- package/dist/skills/opengame-scene-audit/SKILL.md +64 -47
- package/dist/skills/opengame-scene-audit/helpers/scene-audit.js +97 -19
- package/dist/skills/opengame-sdk/SKILL.md +26 -9
- package/dist/skills/opengame-sdk/references/engines.md +2 -2
- package/dist/skills/opengame-sdk/references/identity.md +10 -11
- package/dist/skills/opengame-stability/SKILL.md +59 -27
- package/dist/skills/opengame-trailer-make/SKILL.md +164 -0
- package/dist/skills/opengame-trailer-make/references/archetypes.md +113 -0
- package/dist/skills/opengame-trailer-make/references/briefs.md +273 -0
- package/dist/skills/opengame-trailer-make/references/core.md +129 -0
- package/dist/skills/opengame-trailer-make/references/craft.md +299 -0
- package/dist/skills/opengame-trailer-make/references/edge-cases.md +27 -0
- package/dist/skills/opengame-trailer-make/references/filmable.md +244 -0
- package/dist/skills/opengame-trailer-make/references/finishing.md +126 -0
- package/dist/skills/opengame-trailer-make/references/gates.md +184 -0
- package/dist/skills/opengame-trailer-make/references/measure.md +282 -0
- package/dist/skills/opengame-trailer-make/references/music.md +148 -0
- package/dist/skills/opengame-trailer-make/references/og-clash-record.md +133 -0
- package/dist/skills/opengame-trailer-make/references/ops.md +154 -0
- package/dist/skills/opengame-trailer-make/references/owner-eye.md +214 -0
- package/dist/skills/opengame-trailer-make/references/review.md +141 -0
- package/dist/skills/opengame-trailer-make/references/title.md +92 -0
- package/dist/skills/opengame-trailer-plan/SKILL.md +176 -0
- package/dist/skills/opengame-trailer-plan/references/archetypes.md +113 -0
- package/dist/skills/opengame-trailer-plan/references/briefs.md +273 -0
- package/dist/skills/opengame-trailer-plan/references/core.md +129 -0
- package/dist/skills/opengame-trailer-plan/references/craft.md +299 -0
- package/dist/skills/opengame-trailer-plan/references/edge-cases.md +27 -0
- package/dist/skills/opengame-trailer-plan/references/filmable.md +244 -0
- package/dist/skills/opengame-trailer-plan/references/finishing.md +126 -0
- package/dist/skills/opengame-trailer-plan/references/gates.md +184 -0
- package/dist/skills/opengame-trailer-plan/references/measure.md +282 -0
- package/dist/skills/opengame-trailer-plan/references/music.md +148 -0
- package/dist/skills/opengame-trailer-plan/references/og-clash-record.md +133 -0
- package/dist/skills/opengame-trailer-plan/references/ops.md +154 -0
- package/dist/skills/opengame-trailer-plan/references/owner-eye.md +214 -0
- package/dist/skills/opengame-trailer-plan/references/reference-set.md +92 -0
- package/dist/skills/opengame-trailer-plan/references/review.md +141 -0
- package/dist/skills/opengame-trailer-plan/references/title.md +92 -0
- package/dist/skills/opengame-trailer-review-and-improve/SKILL.md +145 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/archetypes.md +113 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/briefs.md +273 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/core.md +129 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/craft.md +299 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/edge-cases.md +27 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/filmable.md +244 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/finishing.md +126 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/gates.md +184 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/measure.md +282 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/music.md +148 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/og-clash-record.md +133 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/ops.md +154 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/owner-eye.md +214 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/review.md +141 -0
- package/dist/skills/opengame-trailer-review-and-improve/references/title.md +92 -0
- package/dist/skills/opengame-verification/SKILL.md +113 -48
- package/dist/skills/opengame-weapons/SKILL.md +129 -55
- package/dist/skills/opengame-weapons/references/handling-numbers.md +42 -0
- package/dist/skills/opengame-weapons/references/reloads-and-hands.md +96 -0
- package/dist/skills/opengame-weapons/references/sights-and-scopes.md +98 -0
- package/dist/skills/opengame-weapons/references/threejs.md +71 -0
- package/dist/skills/opengame-world-physics/SKILL.md +55 -24
- package/dist/skills/opengame-world-physics/references/measured-failures.md +28 -1
- package/package.json +1 -1
|
@@ -0,0 +1,86 @@
|
|
|
1
|
+
> Carried from `opengame-mobile` 1.1.0. Read it when: Make a browser game run, sound and stay up on real phones. Covers tab memory measured on a real phone in Safari, freeing CPU copies only where nothing re-reads them, texture caps per device class, tier and pixel ratio before the first frame, touch input that keeps every release, audio unlock with a real check, phone frame rates, no blur over the canvas, and a card for a device the game cannot serve. Use when the game crashes, reloads or is silent on a phone, lags on mobile or has stuck touch.
|
|
2
|
+
|
|
3
|
+
|
|
4
|
+
# Mobile
|
|
5
|
+
|
|
6
|
+
A phone is a different machine, and a desktop browser set up to imitate one hides the
|
|
7
|
+
failures that matter. OC (OG Clash) passed every phone gate in Chrome's iPhone emulation
|
|
8
|
+
while the owner's real iPhone crash-looped during load, played no gunfire, and ran at
|
|
9
|
+
10-20 fps. Emulation does not enforce Safari's memory limit, its touch quirks or its
|
|
10
|
+
audio rules. Put a real phone in the loop on the first day.
|
|
11
|
+
|
|
12
|
+
Rules carry five tags; `opengame-quality-bar` explains them and the game codes. Download
|
|
13
|
+
size and caching are in `opengame-loading`; crash counting and recovery in
|
|
14
|
+
`opengame-stability`; frame-time measurement in `opengame-performance`.
|
|
15
|
+
|
|
16
|
+
## Rules
|
|
17
|
+
|
|
18
|
+
### Phone memory is measured on a real phone in real Safari
|
|
19
|
+
- **Invariant:** The game's tab memory on the weakest target phone stays under a written budget through a full session, measured on the device in its own browser, and a build that has passed only emulation has not passed.
|
|
20
|
+
- **Technique:** Connect the phone to a Mac and use Safari's Web Inspector (Timelines, with the Memory instrument) while playing title, menus, a full match and results twice in a row. Add Android Chrome through remote debugging. Treat "A problem repeatedly occurred" or a silent reload as a memory kill. Keep a per-subsystem memory table in the game itself (textures, geometry, audio, WebAssembly heap) so the number has owners. Avoid fixed large up-front memory reservations; let WebAssembly memory grow from a modest start.
|
|
21
|
+
- **Reference configuration:** OC's owner set about 500 MB per tab; it was never met (the desktop tab sat at 3.5 GB, and an iPhone peaked at 3.9 GB). iPhones crash-looped from R165 to R195 after every emulated gate passed. Halving a fixed 4 GiB WebAssembly reservation changed nothing, because only about 0.8 GB of it was the game's own memory. Playwright's WebKit has no WebGPU, so OC had no local iOS reproduction, and fixes were blind until the owner tried their own phone. OS's wasm heap stayed flat while the tab's renderer process grew by 1,444 MB in 9.5 min, so count the whole tab, as `opengame-performance` says; a private window also keeps Cache Storage in memory, so OS capped its cache at 512 MB.
|
|
22
|
+
- **Applies when:** Every game that claims to run on phones, most of all iPhones, where Safari kills tabs over memory with no warning.
|
|
23
|
+
- **Verify:** On the real phone, two full sessions back to back: the tab stays up without a reload, and Web Inspector's peak stays under the budget. Record the phone model and OS version with the result.
|
|
24
|
+
|
|
25
|
+
### CPU-side copies are freed after GPU upload
|
|
26
|
+
- **Invariant:** Once a texture or mesh is on the GPU, the decoded pixels and vertex arrays held for it in JavaScript or WebAssembly memory are released unless something still reads them, and the source URL is kept so they can be fetched again after a context loss.
|
|
27
|
+
- **Technique:** Before releasing a copy, list every reader of it (decals and projected effects, picking and raycasts, CPU-side particles, a later re-upload) and keep the copies those need. After upload, drop references to decoded image data and typed arrays the renderer no longer needs. Keep a table of URLs per asset so a restore can refetch. Three.js: load images with `ImageBitmapLoader`, and with `WebGLRenderer` close the bitmap in `texture.onUpdate` (`texture.image.close()`) once it is uploaded; for meshes you never raycast, morph or rebuild, free each attribute's array after upload with `attribute.onUpload(function () { this.array = null; })`. Test a resize and a forced context loss afterwards, since a texture that must be re-uploaded needs its source again.
|
|
28
|
+
- **Reference configuration:** OC measured about 690 MB of duplicate texture memory, leaving the tab at 4.4 GB ("444 MB is duplicate copies of textures already on the graphics card"), and 180-270 MB on phones. Its runtime had no public way to drop the CPU bytes and keep the GPU texture; a patch of about 30 lines to release them saved about 450 MB. OS released its CPU texture copies the same way and its decals crashed "after about a minute" of play, because the decal system read them; that build was rolled back.
|
|
29
|
+
- **Applies when:** Every textured game on phones; on desktop when tab memory matters.
|
|
30
|
+
- **Verify:** Compare tab memory at the title with release on and off on the real phone, then play a few minutes with every effect that might read a released copy. Force a context loss (`WEBGL_lose_context.loseContext()` and `restoreContext()`): every texture draws again.
|
|
31
|
+
|
|
32
|
+
### Texture size and mips are capped per device class
|
|
33
|
+
- **Invariant:** Each device class has an agreed maximum texture size and downloads only the mip levels it will draw, and any cut in phone fidelity is agreed with the owner before it ships.
|
|
34
|
+
- **Technique:** Pick the device class before loading textures, and point the manifest at per-class files with the top mips trimmed, so a phone downloads only the detail it draws. Keep hero items one step above the rest. Write each cap down with its memory saving and the owner's decision beside it. Three.js: KTX2 files per class through `KTX2Loader`, checked against `renderer.capabilities.maxTextureSize`.
|
|
35
|
+
- **Reference configuration:** OC's phone textures went from 1,063.8 MB to 222.3 MB. The phone cap went from 2048 to 1024 to 512 while crashes continued. Phones had been fetching top mips and discarding them. Unasked phone cuts (16 enemies, 1024 px textures, half-resolution sky) shipped and were reverted after the owner said "i dont want phones to get worse quality".
|
|
36
|
+
- **Applies when:** Every game that ships textures larger than a phone needs.
|
|
37
|
+
- **Verify:** Per device class, sum the texture bytes transferred and resident at the title and in play; the hero items in a phone capture match the desktop capture at phone viewing size.
|
|
38
|
+
|
|
39
|
+
### Tier and pixel ratio are chosen before the first frame
|
|
40
|
+
- **Invariant:** The quality tier, pixel ratio and worker count are set from a device check before the renderer draws anything, so a phone starts at phone settings from its first frame.
|
|
41
|
+
- **Technique:** Classify the device from touch support, screen size, `navigator.hardwareConcurrency`, the GPU adapter and any stored result from an earlier visit, and default phones to the phone tier. Treat a GPU that reports no memory or a generic name as unknown and use the phone tier on a touch device. Set the pixel-ratio cap before the first `render`. Size worker pools from measured speed on the device, not from the reported core count. Log the tier, renderer and pixel ratio the game actually took, and quote that line in every measurement, because a forced weak tier can read the real machine without saying so. Three.js: `renderer.setPixelRatio(Math.min(devicePixelRatio, cap))` before the first frame, and `renderer.compileAsync` for the chosen tier only.
|
|
42
|
+
- **Reference configuration:** OC's phones warmed up at the top tier at full Retina before its caps applied. Safari's GPU reported a generic vendor with 0 bytes and read as desktop, so phones got the top tier. Safari reported 2-4 cores, which would have started 16 threads, so OC fixed workers at 4; on desktop, 8-12 workers ran slower than 4. OS's weak-tier runs used a page shim of `navigator.deviceMemory = 4` that never reached its engine's worker, and the weak-tier switch alone still picked the full renderer from the reported 32 GB, so two days of "weak" numbers ran on the full tier. The pixel budget behind the cap is in `opengame-performance`.
|
|
43
|
+
- **Applies when:** Every game with tiers or a pixel-ratio cap that runs on phones.
|
|
44
|
+
- **Verify:** On the real phone, log the tier and drawing-buffer size of the first frame: both are already the phone values. Force the weak tier on a desktop: the game's own profile line names the weak renderer.
|
|
45
|
+
|
|
46
|
+
### Every touch release arrives, even when the system cancels it
|
|
47
|
+
- **Invariant:** Every touch that starts has an end the game sees, including when the OS cancels it, the finger leaves the canvas or the tab loses focus, so no ghost finger owns the camera and no button stays held.
|
|
48
|
+
- **Technique:** Use pointer events with `setPointerCapture`, and handle `pointerup`, `pointercancel`, `lostpointercapture` and `touchcancel`. Listen for releases on `window`, not on the canvas. Clear every active touch on `visibilitychange` and `blur`. Key touches by identifier, and do not cap the event queue in a way that can drop ends. Set `touch-action: none` on the canvas. Keep the game in touch mode unless a real keyboard or mouse is used for a while.
|
|
49
|
+
- **Reference configuration:** OC's input had no `touchcancel` and a 20-event queue that dropped ends, leaving "a ghost finger that owns the camera forever"; iOS touch ids are unique, so the ghost never cleared, while Chromium emulation hid it. A mouse release heard only on the canvas made 60 of 60 clicks miss after an overlay. One tap flipping the HUD to desktop mode needed a 30-frame guard.
|
|
50
|
+
- **Applies when:** Every game with touch controls, and every game with DOM overlays over the canvas.
|
|
51
|
+
- **Verify:** On the real phone, start a look drag and pull down the notification shade or start the home gesture, return, and drag again: the camera follows the new finger. In automation, send a touch start then a `touchcancel`: no input stays active.
|
|
52
|
+
|
|
53
|
+
### Sound unlocks on the first touch and is checked by measurement
|
|
54
|
+
- **Invariant:** Sound plays from the first touch on a phone, comes back after calls, locks and tab switches, and is proven by measured output on the device, not by the audio context's state or a log line.
|
|
55
|
+
- **Technique:** Create or resume one `AudioContext` inside the first touch handler, and resume again on every touch, on `visibilitychange` and when the context reports `suspended` or `interrupted`, with a short watchdog. On Safari, set `navigator.audioSession.type = 'playback'` before creating the context where the API exists, so the silent switch does not mute the game. Load the player's own weapon sounds first. Give the player's weapon, UI and music priority over enemy voices. Check the output with an `AnalyserNode` on the master bus during a shot, and have a person listen on the phone.
|
|
56
|
+
- **Reference configuration:** OC had no gunfire on iPhone across 3 or more builds, and the owner raised in-game iPhone sound 7 times ("game audio was playing but NOT gunfire sounds"). The phone build's audio functions were stubs while the log still said "unlocked ... 48000 Hz out". Setting the audio session to playback was proposed and its effect was never confirmed. Enemy vocals were capped to the nearest 6 within 20-30 m and the mix to about 40 voices.
|
|
57
|
+
- **Applies when:** Every game with sound that runs on phones.
|
|
58
|
+
- **Verify:** On the real phone with the silent switch on and off: the probe reads a peak on the first shot, after locking and unlocking, and after switching apps and back; a person hears it.
|
|
59
|
+
|
|
60
|
+
### Phones have their own frame-rate floor, measured on the phone
|
|
61
|
+
- **Invariant:** The phone has a written floor and target, measured on the phone in menus, loading and combat, and it holds a steady rate rather than bouncing between two.
|
|
62
|
+
- **Technique:** Budget draw calls and per-enemy costs separately for phones, since phones are bound by draw calls and CPU sooner than by fill. If the phone cannot hold its target, cap at a lower rate it holds evenly. Measure on the device with an on-screen counter and Web Inspector, at the same tier players get. Target the phones most players own, not the oldest or smallest.
|
|
63
|
+
- **Reference configuration:** OC's targets on phones: at least 30 fps, aim 60, with Safari capping animation frames at 60. The owner measured 15-20 fps on their own iPhone ("i dont want to cur quality / fidelity") and OC ended at 10-20 fps at R195. Target devices: iPhone 11 and 12, SE 2 and 3, and mid-range Android ("dont overfit to iphone 5se").
|
|
64
|
+
- **Applies when:** Every game played on phones.
|
|
65
|
+
- **Verify:** On the weakest target phone, record frame times through title, menus, a loading screen and a minute of the heaviest fight, counting frames the game presented rather than animation callbacks (see `opengame-performance`): the one-percent low holds the phone floor.
|
|
66
|
+
|
|
67
|
+
### No backdrop blur over the canvas on iPhone
|
|
68
|
+
- **Invariant:** On iPhone, menus and overlays over the game canvas use pre-blurred or darkened images rather than a live backdrop blur.
|
|
69
|
+
- **Technique:** Replace `backdrop-filter: blur()` over the canvas with a darkened frame, a blurred still captured once, or a solid panel. Keep CSS filters off full-screen layers above WebGL or WebGPU content on phones.
|
|
70
|
+
- **Reference configuration:** OC removed `backdrop-filter` blur over its WebGPU canvas on iPhone, on the suspicion that it cost memory. Nobody measured the memory cost; the cause was a guess.
|
|
71
|
+
- **Applies when:** Every game with DOM menus or HUD panels over the canvas on iOS.
|
|
72
|
+
- **Verify:** On the real phone, compare Web Inspector memory and frame time with the pause menu open and closed; the difference is small.
|
|
73
|
+
|
|
74
|
+
### A device the game cannot serve gets a card or a fallback within a second
|
|
75
|
+
- **Invariant:** Before the large download, the game checks whether this device and browser can run it, and a browser without the needed graphics API, or a touch-only phone opening a game with no touch controls, gets a working fallback or a card that names the reason and what to do, never a blank page or a prompt it cannot answer.
|
|
76
|
+
- **Technique:** Check `navigator.gpu`, then `requestAdapter()` (it can return null even when `navigator.gpu` exists, and Chrome on Linux can expose `navigator.gpu` before an adapter is ready, so request the device yourself with a bounded number of retries inside a deadline, then fall back and record why), then the features and limits the game needs, then WebGL 2 with `getContext('webgl2')` inside a try. Treat the device as touch-only when `matchMedia('(pointer: coarse)')` matches and `(any-pointer: fine)` does not. Decide before fetching the world, and record which backend the game got. If the game falls back, boot the fallback as a real build in a gate (`opengame-stability`). Three.js: `WebGPURenderer` falls back to its WebGL 2 backend when WebGPU is missing (`forceWebGL: true` forces it for testing); check `renderer.backend.isWebGPUBackend` after `await renderer.init()`.
|
|
77
|
+
- **Reference configuration:** OC's capability gate ran in 310-626 ms and named the reason and the fix. Its "No WebGPU" card showed on 60 of 60 boots in testing. OH's launch QA found a phone sat for ever at PRESS ANY KEY and a browser without WebGL 2 showed a blank page; both now show a notice. OS ruled out touch by an owner decision and LC is desktop-only, so for them the card is the whole phone experience. LC's WebGL 2 fallback ran at 44-49 fps against 58-60 on WebGPU, and drew no scattered props for about 11 h before a gate booted it.
|
|
78
|
+
- **Applies when:** Every game built for WebGPU or needing WebGL 2, and every game without touch controls that a phone can open.
|
|
79
|
+
- **Verify:** Boot with WebGPU unavailable (remove `navigator.gpu` in an init script, or a browser without it), and again with WebGL 2 unavailable: the card or the fallback appears in under 1 s, before any large request in the network log. Open the URL on a real phone: a game without touch shows its card.
|
|
80
|
+
|
|
81
|
+
### Touch layout is thumb-sized, landscape and install-aware
|
|
82
|
+
- **Invariant:** Every touch target is large enough for a thumb, the game plays in landscape with a clear prompt in portrait, and players are offered the installed web-app mode without being forced into it.
|
|
83
|
+
- **Technique:** Size targets in points against the safe area, place main actions in the thumb arc, and pause behind a rotate card in portrait. Detect installed mode with `matchMedia('(display-mode: standalone)')` and `navigator.standalone`; show a short per-platform guide to add the game to the home screen with a remembered "play in browser" choice. Declare `display: fullscreen` and `orientation: landscape` in the web app manifest. iOS Safari in a tab has no fullscreen API, no orientation lock and no vibration, so do not depend on them.
|
|
84
|
+
- **Reference configuration:** OC's touch targets are at least 44 pt; its title rows had measured 20-25 px, and some buttons 29 pt. Main buttons sit in the right 30% of the screen, with 12 buttons at most.
|
|
85
|
+
- **Applies when:** Every game with a touch layout.
|
|
86
|
+
- **Verify:** Screenshot every menu and the HUD at a phone landscape size (OC used 844 x 390) and measure each target; play a round on the phone with thumbs only.
|
|
@@ -0,0 +1,183 @@
|
|
|
1
|
+
> Carried from `opengame-montage-make` 1.0.0. Read it when: Build a gameplay montage from its plan, for any game, genre or engine, at any capture tier from a film build to live high-frame-rate capture. Covers the agent team with relays and bounded waits, film-build hooks or replay passes, takes as code with eased speed keys on the beat grid, payoff cinematics from the same run, offline take checks, and a resumable cut chain that delivers read back. Use when a montage plan is ready and it is time to capture and cut.
|
|
2
|
+
|
|
3
|
+
|
|
4
|
+
# Montage make
|
|
5
|
+
|
|
6
|
+
Step 2 of 3: `opengame-montage-plan`, then **make**, then `opengame-montage-review-and-improve` on every round.
|
|
7
|
+
|
|
8
|
+
You build the montage the plan describes. Every take should pass the hard gates before it is offered, and every cut should
|
|
9
|
+
reach the owner's folder without a slate, a placeholder or an unfixed note they were not told about. On the OG Snipers
|
|
10
|
+
sniper montage the rig and film build were up by hour 3, but capture took 42 % of about 302 agent-hours at about 50
|
|
11
|
+
sessions per shipped slot, and 10 h 38 min (16 % of the wall clock) passed with no agent logging: a 7 h 20 min
|
|
12
|
+
model-provider outage and a 3 h 18 min machine stall.
|
|
13
|
+
|
|
14
|
+
## Read first
|
|
15
|
+
|
|
16
|
+
- `plan/PLAN.md`, the common brief and `work/OWNER-NOTES.md` in the film repo.
|
|
17
|
+
- [references/montage-grammar.md](references/montage-grammar.md) (the grammar), [references/montage-capture.md](references/montage-capture.md)
|
|
18
|
+
(tiers, hooks, shooting without source, take checks), [references/montage-genres.md](references/montage-genres.md)
|
|
19
|
+
(fake tells, probes and HUD per genre), [references/montage-ops.md](references/montage-ops.md) (team, relay, slots,
|
|
20
|
+
the cut chain, the time budget) and [references/montage-gates.md](references/montage-gates.md).
|
|
21
|
+
- [references/filmable.md](references/filmable.md) (the capability checklist, stepped capture, the per-engine table,
|
|
22
|
+
shot specs as code, the animation audit), [references/finishing.md](references/finishing.md) (conform, render, QA,
|
|
23
|
+
delivery) and [references/edge-cases.md](references/edge-cases.md) (closed builds, 2D, VR, no GPU). `opengame-capture`
|
|
24
|
+
has the in-game hook rules if you own the game.
|
|
25
|
+
- [references/song.md](references/song.md), [references/reference-study.md](references/reference-study.md) (the
|
|
26
|
+
finishing layer, the bot tells and the player model), [references/briefs.md](references/briefs.md) (common, shooter,
|
|
27
|
+
resume and sweep briefs), [references/owner-eye.md](references/owner-eye.md),
|
|
28
|
+
[references/measure.md](references/measure.md) (how to measure each check with ffmpeg and Python),
|
|
29
|
+
[references/day-one-questions.md](references/day-one-questions.md) (what the owner settled on day one) and
|
|
30
|
+
[references/og-snipers-record.md](references/og-snipers-record.md).
|
|
31
|
+
- Rename the shooter words for your genre: payoff (kill, goal, overtake, KO, clear), receiver (victim, overtaken car,
|
|
32
|
+
keeper, losing team, board), secondary sounds (reload, gear change, footwork).
|
|
33
|
+
|
|
34
|
+
## 1. Team and ops (the first 30 minutes)
|
|
35
|
+
|
|
36
|
+
- **Roles:** a director; a film-build engineer and a rig engineer in parallel against one spec contract; shooters, one per
|
|
37
|
+
range of slots; edit; sound; post; ending; an independent sweep per cut. One owner per file set. The director makes no
|
|
38
|
+
captures once shooters exist. About 10 agents ran at once on OG Snipers.
|
|
39
|
+
- **Relay discipline.** After every owner message: rows in the notes ledger and `work/OWNER-NOTES.md`, then a relay
|
|
40
|
+
checklist of every role it touches, written to each `inbox/<role>.md` and the shared bus. Every agent reads its inbox
|
|
41
|
+
at the start of every task and before queuing any take; specs are read at session start so queued sessions pick up late
|
|
42
|
+
notes. (A title hand-in finished before the note that changed it reached that agent; shooters queued takes a newer note
|
|
43
|
+
had cancelled.) Watch shared files with `tail -n 0 -F`.
|
|
44
|
+
- **Handshakes** through `status/<role>.json` (author, cut id, commit, READY), never by grepping a shared log.
|
|
45
|
+
- **Bounded waits:** every wait ends on success, on a failure line, or after 15 minutes, then reads the real state.
|
|
46
|
+
**Liveness:** each heartbeat lists every agent's last activity; more than 30 minutes silent with no job of its own means
|
|
47
|
+
hung: restart it from its resume brief. A heartbeat skips if the last one has not finished; delete every schedule at
|
|
48
|
+
delivery. The director's own turn stays alive through long waits; a context restart resumes from the resume brief and
|
|
49
|
+
`plan/DECISIONS.md`.
|
|
50
|
+
- **Provider failover:** resume briefs on disk at all times; at the first wave of refusals, restart every agent on another
|
|
51
|
+
model provider within 15 minutes, among those the person's setup already has (never buy credit without asking).
|
|
52
|
+
- **The machine:** agree the share before production and verify it is live in the tool everyone calls. Use your machine's
|
|
53
|
+
job queue if it has one: a fixed slot count with timed raises that revert; takes ahead of other jobs ahead of proofs;
|
|
54
|
+
tickets older than 40 minutes jump ahead; one reserved slot for the director's finishing passes; withdraw only waiting
|
|
55
|
+
tickets, by verified process id. Pack takes into one session (OG Snipers booted in 130-180 s), sessions under 12
|
|
56
|
+
minutes. **Measure and report the queue's arrival and service rates** (OG Snipers: 14 takes an hour arriving, 8 served)
|
|
57
|
+
and cut work that fails preflight before it queues. Measure how many parallel captures your machine holds (OG Snipers:
|
|
58
|
+
three held, four lost takes to GPU device loss).
|
|
59
|
+
- **Disk:** build the delete keep-list from the select files at deletion time; caps per role; archive selects off the
|
|
60
|
+
capture machine.
|
|
61
|
+
|
|
62
|
+
## 2. Build the set
|
|
63
|
+
|
|
64
|
+
### The film build (tier A: one owner, a private worktree)
|
|
65
|
+
|
|
66
|
+
Fork the frozen source; every switch's default equals shipped, proved with a parity frame: a per-frame command bridge with
|
|
67
|
+
a far keepalive; a stepped clock with any step and a slow floor of 0.05x or lower; **a seed over every random generator**;
|
|
68
|
+
HUD per element and event filters (a hero-only feed, payoff badges on their own frame); a stage camera with a history cut
|
|
69
|
+
and attach spaces; every body drawn in sync each frame under slow motion; opponent or AI staging (place, act, path,
|
|
70
|
+
loadout, name, health or pace); the per-frame probe (player, camera, every actor, every payoff event with receiver and
|
|
71
|
+
cause, audio cues) **plus, on day one, the drawn points of whatever the engine animates worst (a skeleton, wheels, board
|
|
72
|
+
pieces, units), the payoff point, the line of the attack, ground height and a camera look-at** (they came on day 3 on OG
|
|
73
|
+
Snipes, after two guessed framings failed); an input assist inside the game's own rules (OG Snipers: an aim trim of at most
|
|
74
|
+
0.35° before the game's hit test); reaction controls (ragdoll, crash, KO, a hitstop multiplier); a shutter in game time.
|
|
75
|
+
Put film switches on run-time structs (a field on a serialised struct broke two builds); staging acts on a change of value
|
|
76
|
+
(send a reset 6 frames before). A proof table per hook in the hooks file.
|
|
77
|
+
|
|
78
|
+
### Without source (tiers B-E)
|
|
79
|
+
|
|
80
|
+
No film build: say what that removes and plan around it (the capture chapter). B: record a human's input live, re-run it
|
|
81
|
+
in tier A's rig, hash state per frame to prove the replay. C: play live, then shoot each moment in passes from the replay
|
|
82
|
+
(as played, outside, the receiver's side), rendering offline if the replay allows; check its slow motion on frames for
|
|
83
|
+
interpolation (use interpolated replays only to 0.5x); version-lock the build. D: the game's own cameras plus
|
|
84
|
+
high-frame-rate capture. E: lossless capture at the highest stable rate (the slow-motion budget is capture fps ÷ 60),
|
|
85
|
+
events mined from the feed or score pops, no second angle. The rest of this skill holds; staging and seeds become
|
|
86
|
+
selection from many real plays.
|
|
87
|
+
|
|
88
|
+
### The rig
|
|
89
|
+
|
|
90
|
+
The stepped capture for the tier (the filmable chapter); seeded replay proved on day one; **one shared clock library with a
|
|
91
|
+
unit test against the engine's own time** (a clock-sum off-by-one between the payoff solver and the engine failed one take
|
|
92
|
+
three times); a human-input synthesiser inside the measured ranges of a good player of this genre (the reference-study
|
|
93
|
+
chapter); a payoff solver that lands payoffs only on grid frames; the speed logged on every frame; a renderer check that
|
|
94
|
+
refuses software rendering; take checks (frame count, duplicate signatures, clock skips, quality, every payoff the spec
|
|
95
|
+
asked for actually happened).
|
|
96
|
+
|
|
97
|
+
### Audits and the pilot
|
|
98
|
+
|
|
99
|
+
Audit every reaction, hero move and opponent behaviour on frames before shooting on it; scout every map or level for
|
|
100
|
+
z-fighting and flicker at the planned lenses (change the framing or the place; a film-only backdrop fix did not cure one);
|
|
101
|
+
send the skin and loadout casting sheet; measure the genre's fake tells. Shoot the plan's **pilot clip** first and send it.
|
|
102
|
+
|
|
103
|
+
## 3. Shoot
|
|
104
|
+
|
|
105
|
+
- **Specs as code** from `grid.json`: payoffs on grid frames at the beat − 3 frames; movement accents with measured leads;
|
|
106
|
+
speed keys eased per AE1-AE4; the frozen spec, seed and build saved with every select.
|
|
107
|
+
- **Preflight offline before a take takes a slot:** every way the scripted run can fail, checked without a GPU. On OG
|
|
108
|
+
Snipes each of these lost real takes: lines of fire, thin cover, shields, lobby bots, pace marks on another payoff's
|
|
109
|
+
line, clock sums. Racing: the line, AI pace, contacts. Fighting: inputs against frame data.
|
|
110
|
+
- **Clock-easing traps:** changing the clock moves where an action lands (an eased clock after a hop lost a kill); at 0.3x
|
|
111
|
+
one simulation frame spans 3.3 output frames, so a key may be unreachable; a payoff cinematic under 28 frames cannot fit
|
|
112
|
+
the V ramp. Check every take's speed log: fail a one-frame change over 15 % and a constant run of 20 frames or more after
|
|
113
|
+
a payoff.
|
|
114
|
+
- **The payoff cinematic comes from the same run:** the seeded take with a stage camera, the input replay, or the replay's
|
|
115
|
+
free camera on the same moment. Cut on the payoff + 1 frame, from the receiver's side, under the grammar chapter's rules.
|
|
116
|
+
**Solve the lens in two passes** when the framing must include X and exclude Y on every frame: pass 1 with any lens and
|
|
117
|
+
the probe logging; solve a per-frame lens with an 8 % margin; pass 2 plays the keys on the same seed, build and clock;
|
|
118
|
+
check pass 2's own logged points; if the receiver inverts, cut to close-ups or re-stage. OG Snipers verified 0 of 55, 52
|
|
119
|
+
and 42 frames with a leg point in frame.
|
|
120
|
+
- **The take check before offering** (about 20 s, no GPU): the grid, no dead time, only the high-skill form, opponents
|
|
121
|
+
moving, HUD and feed names right, cause before effect, the receiver matching, smoothness on every frame. Offer with three
|
|
122
|
+
full-size graded frames, an every-frame strip of every receiver and hero shot, and what is weak, in so many words.
|
|
123
|
+
- **Never cut a shot the owner liked:** fix it by re-staging, re-framing or a build request. Never delete a master a select
|
|
124
|
+
names.
|
|
125
|
+
- **The master format:** capture at it, or keep every select seeded and re-shoot it at the master size, verified (frame
|
|
126
|
+
count, speed curve, payoff frames, receivers, the camera within 0.15 m and 4 % field of view) before the master.
|
|
127
|
+
|
|
128
|
+
## 4. Cut, mix, finish, deliver: one chain per cut
|
|
129
|
+
|
|
130
|
+
- **The edit list never retimes:** no speed, hold, blend, fade or blur keys. At tiers C-E the conform that lays
|
|
131
|
+
high-frame-rate footage at 60 fps is the one sanctioned rate change, declared per shot. A rolling working cut follows
|
|
132
|
+
select changes (at most once every 10 minutes under load) so a full-length cut always exists.
|
|
133
|
+
- **At the director's word the edit list freezes** and a resumable chain runs it, each stage idempotent and restartable
|
|
134
|
+
by name (one OG Snipers cut needed five resume scripts written by hand before the chain existed). Never edit a running
|
|
135
|
+
script: write a temporary file, then rename it.
|
|
136
|
+
1. **Lint take-bound rows:** every repair, reframe, push and look exception names its take; stop with an error when its slot's
|
|
137
|
+
take changed, and re-bind in the same order that rules the new take in. (A z-fight repair stopped applying without any warning
|
|
138
|
+
for five cuts while the owner was told those cuts were checked; a title was graded twice.)
|
|
139
|
+
2. Render; **scan for flicker only in the spans that changed,** verdict before strips, a cleared list by take.
|
|
140
|
+
3. **One finishing pass** on the reserved slot, against the locked approved look; frames in equal frames out; the title
|
|
141
|
+
box against its approved comp.
|
|
142
|
+
4. **Mix by signature:** the mix carries the edit list's signature and is muxed only on a match; reuse an existing mix
|
|
143
|
+
whose signature matches. Wait for the sound agent's READY at most 40 minutes, then hold and say so.
|
|
144
|
+
5. Remux bit for bit; QA on the final file.
|
|
145
|
+
6. **Deliver to the owner's folder and check it landed:** a plain temporary name (never a name starting with a dot),
|
|
146
|
+
rename, read it back and compare size and checksum, retry after 10 minutes, report "not delivered" until it is
|
|
147
|
+
there. The copy is a chain stage, not a habit: twelve OG Snipers cuts never reached the owner's folder.
|
|
148
|
+
7. **The full gate report after delivery,** as a delta (new FAIL, fixed, standing); the hard gates and the flicker hold
|
|
149
|
+
already ran before the send. The independent sweep runs on the edit render while stages 3-5 run.
|
|
150
|
+
- **Sound** is one command inside every cut build: the song untouched, payoffs timed by their attack, levelled as heard to
|
|
151
|
+
one target, secondary sounds at least 6 dB under, full tails, measured again on the delivered file.
|
|
152
|
+
|
|
153
|
+
## 5. Before anything is sent
|
|
154
|
+
|
|
155
|
+
- Watch the cut once at 1x, start to finish, and measure its sound (the OG Snipers director never watched; the owner always
|
|
156
|
+
did).
|
|
157
|
+
- The hard gates pass or are listed; every `OWNER-NOTES.md` row verified on frames of the delivered file, with NOT VERIFIED
|
|
158
|
+
rows at the top of the message; the sweep landed, or the cut goes out "sweep pending" with its A rows within the hour.
|
|
159
|
+
- No slates, no placeholders, no proxies. Then run `opengame-montage-review-and-improve`, and add a line to
|
|
160
|
+
`film/SKILLS.log`.
|
|
161
|
+
|
|
162
|
+
## The traps that cost the most on OG Snipers
|
|
163
|
+
|
|
164
|
+
Faulted shots cut and filled with shots tied to nothing: about 12-14 hours. One model provider: 7 h 20 min dead. The
|
|
165
|
+
payoff framing decided on day 3: about 10-12 hours. The hero cast without asking: about 10 hours. The title and ending
|
|
166
|
+
without stills: 26 agent-hours and 23 hand-ins. An open wait on a dead job: 3.6 hours; an unnoticed idle agent: 4.4 hours.
|
|
167
|
+
Slot queues: a proof starved 2 h 20 min; a neighbour borrowed slots for 4 hours. A placeholder direction check: one round.
|
|
168
|
+
A voice line untested over the song: six episodes and three masters. A take-bound repair that stopped applying: five cuts
|
|
169
|
+
with a known z-fight. Waiting inside a cut: 45 minutes for the sound agent. Spacing copied from the reference: two re-shoot
|
|
170
|
+
rounds. Capture at 1080p for a 4K master: an upscale. Masters of 11 takes deleted while selects named them.
|
|
171
|
+
|
|
172
|
+
## Don't
|
|
173
|
+
|
|
174
|
+
- Don't retime, hold, blend or fade in the edit; speed lives in the capture (or in the capture's own frame rate).
|
|
175
|
+
- Don't queue a take without reading your inbox and preflighting it offline.
|
|
176
|
+
- Don't wait without a timeout, or signal readiness through a shared log.
|
|
177
|
+
- Don't let a take-bound row survive a take swap unlinted.
|
|
178
|
+
- Don't mux a mix whose signature does not match the edit list.
|
|
179
|
+
- Don't run more than one finishing pass per cut, or grade an approved element twice.
|
|
180
|
+
- Don't claim a delivery you have not read back from the owner's folder.
|
|
181
|
+
- Don't hold a cut the owner is waiting for on the full gate report.
|
|
182
|
+
- Don't edit the game team's source; send fixes through the game-notes file.
|
|
183
|
+
- Don't spend on generated sound or voice without the person's go-ahead, on their own account.
|
|
@@ -0,0 +1,293 @@
|
|
|
1
|
+
# Montage briefs: what to hand each agent
|
|
2
|
+
|
|
3
|
+
Write a brief to disk before spawning any agent; the brief is the agent's whole context. Fill every `{brace}` and paste
|
|
4
|
+
the whole brief. These are the briefs the OG Snipers montage ran on, with what each one learned.
|
|
5
|
+
|
|
6
|
+
## 1. Common brief (every agent reads it first)
|
|
7
|
+
|
|
8
|
+
> Film: a {length} s gameplay montage of **{GAME}** ({genre, engine}), cut to {the song, BPM}, in the style of {the
|
|
9
|
+
> reference edit}. Master {format}; review cuts at 1080p60. The song is played, not edited; the music leads the mix.
|
|
10
|
+
> Frozen build {id, hash}; film build {path}, owned by {role}; capture tier {A-E}. The owner's words are in
|
|
11
|
+
> `work/OWNER-NOTES.md` and the notes ledger; the praised list is `work/PRAISED.md`.
|
|
12
|
+
>
|
|
13
|
+
> House rules: write only inside {your output path}; never modify the game team's workspace; never kill a process you did
|
|
14
|
+
> not start (by verified process id, never by pattern). Heavy jobs (browsers, captures, renders, encodes) go through
|
|
15
|
+
> {your machine's job queue, if it has one}, with thread caps. Never play audio aloud. Never print or log keys, or pass them as command-line
|
|
16
|
+
> arguments; read them inside the process. Never use `git stash` in a shared checkout. Keep {N} GB of disk free and
|
|
17
|
+
> check before every take; delete rejected masters only from a keep-list built from the select files. Requests to
|
|
18
|
+
> another agent go in `work/REQUESTS.md`; decisions go in `plan/DECISIONS.md`, one dated line each. Paid generation
|
|
19
|
+
> only on the person's own account, through the capped ledger; approval: {verbatim}. Read your inbox
|
|
20
|
+
> (`inbox/{role}.md`) at the start of every task and before queuing any take. Report readiness in
|
|
21
|
+
> `status/{role}.json` (author, cut id, commit, READY), never only in a log. Every wait ends on success, a failure line,
|
|
22
|
+
> or 15 minutes, then reads the real state. A progress line every 20 minutes with the UTC time from `date -u`. Report
|
|
23
|
+
> measurements, not impressions; never invent a number; quote the owner, never paraphrase them.
|
|
24
|
+
|
|
25
|
+
## 2. Research fan-out (plan step; on OG Snipers seven agents landed in 58 minutes)
|
|
26
|
+
|
|
27
|
+
Every research agent is read-only, writes one file into `research/`, backs every claim with a file and line or a
|
|
28
|
+
measurement, and writes a progress line every 20 minutes.
|
|
29
|
+
|
|
30
|
+
- **R-GAME:** the game bible (the filmable chapter): the money event from score events, announcer lines and banners; its
|
|
31
|
+
read time; whether the game has replay, determinism and hitstop; the HUD that carries the stakes; the genre's fake tell;
|
|
32
|
+
known risks to footage.
|
|
33
|
+
- **R-CAPTURE:** the capability checklist C1-C19 and the capture tier, with one 20-minute probe judged on frames (replay
|
|
34
|
+
fidelity, slow-motion duplicates, ragdoll or physics stepping, HUD toggles, debug overlays).
|
|
35
|
+
- **R-MUSIC:** measure the song into `research/grid.json` (the song chapter): tempo and drift, bar 1, sections, the vocal's
|
|
36
|
+
end, the door to the end shot.
|
|
37
|
+
- **R-REF:** the reference edit measured as a study (the reference-study chapter), with a rubric for the judges and a
|
|
38
|
+
steal list.
|
|
39
|
+
- **R-REFS:** 3-5 more edits of this genre, measured (cuts, shot lengths, beat lock, ramp shape, flashes, grade).
|
|
40
|
+
- **R-SOUND:** the game's sound inventory organised by payoff, movement and secondary sounds; its callouts and their
|
|
41
|
+
intelligibility over the song.
|
|
42
|
+
- **R-PLAYER:** 3-5 minutes of a good human playing the real game, measured into a timing and aim model (acquisition,
|
|
43
|
+
settle, hold, drift) with bot-tell thresholds (the reference-study chapter shows the sniping version).
|
|
44
|
+
|
|
45
|
+
Never let research pick the hero, the skins, the maps or the title: those are the owner's taste, asked on day one.
|
|
46
|
+
|
|
47
|
+
## 3. Shooter (one per range of slots)
|
|
48
|
+
|
|
49
|
+
> Read first: the common brief, your inbox, `plan/PLAN.md` (the slot table, the payoff model, the speed plan),
|
|
50
|
+
> `plan/shots.json` (your slots), `research/grid.json`, the film build's hooks file, the grammar chapter, the owner-notes
|
|
51
|
+
> table and the praised list.
|
|
52
|
+
>
|
|
53
|
+
> For each slot: write the spec as code from `grid.json` (payoffs on grid frames at the beat − 3 frames; movement accents
|
|
54
|
+
> with their measured leads; speed keys eased per AE1-AE4; the seed, build and clock frozen with the select). Preflight it
|
|
55
|
+
> offline before it takes a slot (lines of fire, cover, opponents' paths, clock sums). Shoot; run the take checks;
|
|
56
|
+
> shoot the payoff cinematic from the same run (the seeded take with a stage camera, the input replay, or the replay's
|
|
57
|
+
> free camera), with the two-pass lens solve when the framing must include X and exclude Y on every frame.
|
|
58
|
+
>
|
|
59
|
+
> Offer a take only with three full-size graded frames, an every-frame strip of every receiver and hero shot, the take
|
|
60
|
+
> check's results, and what is weak, said plainly. Never cut a shot the owner praised: fix it by re-staging, re-framing or
|
|
61
|
+
> a build request. Never delete a master a select names. Self-review every take harshly: three honest lines (great,
|
|
62
|
+
> weak, with another hour). Keep at most one master-resolution take in flight, in your own port range. Report in
|
|
63
|
+
> 120-300 words: per slot the select, payoff accuracy, the take check, what is placeholder, what is unchecked, and the
|
|
64
|
+
> calls the owner might want to change.
|
|
65
|
+
|
|
66
|
+
## 4. Resume brief (on disk at all times)
|
|
67
|
+
|
|
68
|
+
> Resume brief #{n} ({role}, {UTC time}). The previous session {hit a usage limit / lost its provider / restarted}. Read
|
|
69
|
+
> this, the common brief, your inbox, `plan/DECISIONS.md` and your role's last lines in the progress log. Do not redo what
|
|
70
|
+
> exists: check select timestamps, take check files and proxies first. What changed since the stop: {the owner's newest
|
|
71
|
+
> notes verbatim; builds; machine share}. Your items and their state: {list}. Unreviewed takes that landed after the stop:
|
|
72
|
+
> {review, select or reject; do not reshoot}. Scope, if frozen: {item, state, owner}; if one item runs more than about
|
|
73
|
+
> 45 minutes behind, cut without it and say which. Finish line: {the next cut and what it must carry, then the sweep,
|
|
74
|
+
> the A/B against the last cut, delivery to the owner's folder read back; the master only when the owner says final}.
|
|
75
|
+
> When done, write `{role} FINAL DONE` in the progress log.
|
|
76
|
+
|
|
77
|
+
## 5. The independent every-frame sweep
|
|
78
|
+
|
|
79
|
+
For the director (not pasted): the review source with the best hit rate on OG Snipers. It flagged 14.5 of the 29 owner
|
|
80
|
+
defect notes it could have seen (50 %), against 10 % for blind critics and gates and 0 % for a judge panel that never
|
|
81
|
+
ran. Its independence is the point: a sweep that read no earlier report or ruling found an alternate-frame head flip the
|
|
82
|
+
director had ruled in. Run one on every cut that will be sent, on the edit render while the finishing pass runs, so it
|
|
83
|
+
lands **before** the send. Use a different model each time, rotating vendors among those the person's setup already has.
|
|
84
|
+
Check a sample of its A rows on the delivered file before trusting it, and write TRUE, OVERSTATED, FALSE or DESIGN beside
|
|
85
|
+
every row you act on. It reports; it never decides. It once recommended cutting every faulted shot the owner had praised,
|
|
86
|
+
and following it cost about a day: fill in the praised list every time.
|
|
87
|
+
|
|
88
|
+
> You are sweeping a gameplay montage cut frame by frame for everything that looks cheap, broken, fake or unfinished. The
|
|
89
|
+
> owner has said: "im not going to call out every single slop point u need to find all this urself". Your job is to be
|
|
90
|
+
> the eye that finds every such point before they do. Be exhaustive and exact; no praise, no taste essays, no scores.
|
|
91
|
+
>
|
|
92
|
+
> Read only, on every project file. Write only your report and your own scratch under {scratch dir}. Do the work
|
|
93
|
+
> yourself; start no subagents. The machine is shared: `nice -n 10`, ffmpeg `-threads 6`, no GPU, no browser, never play
|
|
94
|
+
> audio aloud. Keep scratch under {2 GB}; delete each frame dump once you have looked at it.
|
|
95
|
+
>
|
|
96
|
+
> THE FILM: {cut path} ({N} s, {frames} frames at {fps}, {resolution}): a montage of {GAME, one line}, cut to {the song,
|
|
97
|
+
> BPM}. It has {primary-view clips} with outside cinematics cut in ({what they show}), an opening and an end shot {with /
|
|
98
|
+
> without a title}. Do NOT read the project's plans, notes, reviews, rulings, earlier sweeps or code. You may read
|
|
99
|
+
> {ledger lines} (the owner's words) and the reference edit {path}.
|
|
100
|
+
>
|
|
101
|
+
> WHAT THE OWNER HAS ALREADY CALLED SLOP (their words; find every further instance of each, and anything of the same
|
|
102
|
+
> family they have not named): {their complaint lines, verbatim}. The families from earlier films, which hold for any
|
|
103
|
+
> genre:
|
|
104
|
+
> - surface flicker or z-fighting anywhere on a map or level;
|
|
105
|
+
> - slow motion that judders, steps or drops frames on any object; a character jittering or shaking;
|
|
106
|
+
> - an opponent or AI stuck, running into geometry, shuttling back and forth, idle, or in a row facing the lens; clones;
|
|
107
|
+
> - any part passing through another: limbs through stairs, rails or ground, a weapon through a body or a wall;
|
|
108
|
+
> - a reaction shown before its cause; the same payoff shown twice;
|
|
109
|
+
> - an outside shot showing the player-view rig (a floating arm) or the wrong subject, place or moment for its payoff; a
|
|
110
|
+
> hero shot that crops the character the owner wanted seen whole;
|
|
111
|
+
> - menus, loading screens, lobbies or results screens anywhere in the film;
|
|
112
|
+
> - an angle where the payoff cannot be read; an insert in the middle of a chain with no reason;
|
|
113
|
+
> - dead time after a clip's last payoff; the player searching (looking around) between payoffs;
|
|
114
|
+
> - the low-skill form of the core action where only the high-skill form is allowed; a trick at slowed speed;
|
|
115
|
+
> - {the parts of a body the owner wants out of frame on impact shots};
|
|
116
|
+
> - HUD that differs between clips (spacing, anchor, icons), names repeated in a feed, HUD the owner did not allow;
|
|
117
|
+
> - a speed change in one frame; a payoff cinematic at one constant slow speed; a held or duplicated frame; a stray frame;
|
|
118
|
+
> - a sound louder than the payoffs, a payoff noticeably louder or quieter than the rest, a secondary sound over its
|
|
119
|
+
> payoff, a tail cut dead, a stray tone or beep, the song changed or dropping out;
|
|
120
|
+
> - the genre's own fake tells: {from the genre chapter}.
|
|
121
|
+
>
|
|
122
|
+
> PRAISED: {the shots the owner said they like, with frames and their words}. On a praised shot report ONLY the families
|
|
123
|
+
> above that the owner themselves complained about, as fixes, and mark every other remark on it "taste, not a defect".
|
|
124
|
+
> Never recommend CUT for a praised shot: recommend the smallest fix.
|
|
125
|
+
>
|
|
126
|
+
> METHOD.
|
|
127
|
+
> 1. Extract EVERY frame (in ranges of about 300) and look at them with your image tool: every outside, cinematic, hero
|
|
128
|
+
> and payoff shot on every frame at full size; the primary view on every 3rd frame plus every frame from 14 before to
|
|
129
|
+
> 20 after each payoff. Make tiled strips of consecutive frames (6-8 a row) so motion, stutter and flicker show across
|
|
130
|
+
> neighbours. Never judge a body from every-2nd-frame sheets: alternate-frame flips hide there.
|
|
131
|
+
> 2. Measure as well as look: per-frame difference of a tracked region; a body that does not move on a frame while its
|
|
132
|
+
> neighbours do; a small region alternating A-B-A-B (on head, hand and weapon crops, not only the hips); a one-frame
|
|
133
|
+
> camera jump.
|
|
134
|
+
> 3. Follow every opponent who is ever in frame: where are they going, are they blocked, idle, sliding, duplicated,
|
|
135
|
+
> popping?
|
|
136
|
+
> 4. Put the same crop of every recurring HUD element from every clip on one sheet and diff the anchors in pixels.
|
|
137
|
+
> 5. Sound, on the delivered mix: every payoff's in-band RMS (150 Hz-8 kHz over 300 ms from its attack) in one table with
|
|
138
|
+
> the median and spread; every transient ranked (name anything louder than the loudest payoff); secondary sounds
|
|
139
|
+
> against their clip's payoff; callout offsets and levels; a spectrogram pass for stray tones.
|
|
140
|
+
>
|
|
141
|
+
> REPORT {report path}:
|
|
142
|
+
> 1. One table, one row per defect, sorted by frame: frames (first to last), clip or shot, family, exactly what is wrong
|
|
143
|
+
> and where in the frame, severity (A: the owner would stop on it; B: visible on a second watch; C: frame-step only),
|
|
144
|
+
> the smallest fix (re-frame, re-stage, trim to frames x-y, re-shoot, level by N dB).
|
|
145
|
+
> 2. Standing asks: each of the owner's rules {list} with PASS or FAIL and the evidence.
|
|
146
|
+
> 3. The ten worst, in order.
|
|
147
|
+
> 4. Every outside shot marked KEEP, FIX or CUT (never CUT a praised shot), and the shots you found clean.
|
|
148
|
+
> 5. How many frames you looked at, and anything you could not check.
|
|
149
|
+
>
|
|
150
|
+
> Append a progress line at the start and at the finish. Final message: counts by family and severity, the ten worst
|
|
151
|
+
> with frames, and the standing asks that FAIL.
|
|
152
|
+
|
|
153
|
+
## 6. The judge panel: three blind judges against the golden edit
|
|
154
|
+
|
|
155
|
+
For the director (not pasted):
|
|
156
|
+
|
|
157
|
+
- **Why it exists.** The OG Snipers owner asked for "a council/panel of judges ... each give scores ... where 10/10 is the
|
|
158
|
+
level of [the reference edit] ... and score our edit across diff dimensions out of 10 vs the golden set". It never ran:
|
|
159
|
+
it was held for "the bar is met" at every cut from the third to the nineteenth, and the owner called the final with no
|
|
160
|
+
score.
|
|
161
|
+
- **Who.** Three independent judges from different model vendors (the owner's named models if they give them; whatever
|
|
162
|
+
the person's agent setup already has, never bought credit without asking), on high effort, each blind to plans, code,
|
|
163
|
+
notes and each other. The same brief and rubric every round, so scores move only because the film moved.
|
|
164
|
+
- **When.** A baseline on the first full cut that passes QA and the owner-note gates; then every third delivered cut, or
|
|
165
|
+
as soon as a cut passes the director's own bar, or after three cuts with no new owner note; on the final candidate;
|
|
166
|
+
whenever the owner asks. If the owner says not to start it until the bar is met, run it anyway as a measuring stick and
|
|
167
|
+
keep its verdict out of their updates until then; say so once.
|
|
168
|
+
- **Reading it.** Each dimension's median and spread; a spread over 2 points means measure it, not argue it. Map every
|
|
169
|
+
owner note to a dimension; where the judges scored 7 or more and the owner complained, rewrite that rubric line in the
|
|
170
|
+
owner's words. Treat the gap list like any critic's: filter it through the ledger and the montage overrides, reject any
|
|
171
|
+
fix that contradicts the owner (designed silences, varied gaps, shorter tails, held scopes and cutting praised shots
|
|
172
|
+
all did on OG Snipers), verify the rest on frames of the delivered file, then route. A judge reports; it never decides.
|
|
173
|
+
- **The finish line** is the owner's approval unless they name the panel's target as the finish line.
|
|
174
|
+
|
|
175
|
+
> You are one of three independent judges on a panel. You are a veteran of two worlds: the editor who cuts the best
|
|
176
|
+
> {genre} montages ({the reference editor}) and the creative director of {the genre's official gameplay trailers}. You did
|
|
177
|
+
> not make the film you are judging and have no stake in it. Be severe and exact. A score is evidence, not
|
|
178
|
+
> encouragement: praise only what is truly at the bar. You judge alone; do not look for the other judges' reports.
|
|
179
|
+
>
|
|
180
|
+
> Read only. Do not edit any project file. Do not read the film's source code, plans, briefs, notes, review files,
|
|
181
|
+
> rulings or history: judge only what is on screen and in the sound. Never play audio aloud; measure it.
|
|
182
|
+
>
|
|
183
|
+
> THE FILMS. GOLD (10/10 by definition): {path}, {editor, title, length, rate}. The owner's words on it: {verbatim}. Its
|
|
184
|
+
> measured study: {research/REF-STUDY.md} (read it whole). If GOLD has no ending, use {other edits by the same editor} for
|
|
185
|
+
> endings. OURS: {cut path} ({resolution}, {rate}, {N} s): a montage of {GAME}, cut to {the song}. The song file alone:
|
|
186
|
+
> {path} (for a null test). Its gate report {path} may be used and should be spot-checked.
|
|
187
|
+
>
|
|
188
|
+
> WHAT THE OWNER ASKED FOR (do not mark the film down for obeying them): {their standing rules for this film, verbatim}.
|
|
189
|
+
> Never award points for a device of GOLD that they banned, and never subtract them for its absence.
|
|
190
|
+
>
|
|
191
|
+
> METHOD. ffmpeg and Python only (`nice -n 10`, `-threads 6`), a private environment under {scratch dir}, raw output there
|
|
192
|
+
> (under {3 GB}); write nothing else except your report. Watch OURS start to finish once as a viewer would (a 12 fps pass
|
|
193
|
+
> in order) before taking it apart. Look at frames: 2 fps overviews of both films; every-frame strips across every cut,
|
|
194
|
+
> ramp, transition and payoff in OURS, and the matching moments in GOLD. Frame-exact grabs with `select=eq(n\,N)`, never `-ss`.
|
|
195
|
+
> Measure: cuts and shot lengths; payoff spacing in beats; payoffs on the beat at the picture lead; the payoff cycle
|
|
196
|
+
> (glide, rise, peak, fall) and its peak over glide; seam speeds in and out; duplicate and held frames; luma per section
|
|
197
|
+
> and at cuts; loudness; every payoff's in-band RMS and its spread; any transient louder than the loudest payoff; the
|
|
198
|
+
> song against the supplied file (untouched?). "Payoff" is the genre's peak event: {kill / goal / overtake / combo ender /
|
|
199
|
+
> boss break / clear}.
|
|
200
|
+
>
|
|
201
|
+
> THE RUBRIC: score each 0-10, GOLD = 10, with a timestamp in OURS, the matching moment in GOLD, and a number where one
|
|
202
|
+
> exists. Half points allowed; above 10 with a reason; a dimension where OURS equals GOLD is a 10.
|
|
203
|
+
>
|
|
204
|
+
> | # | Dimension | What 10 looks like | Evidence required |
|
|
205
|
+
> |---|---|---|---|
|
|
206
|
+
> | 1 | Sync to music | every payoff on the beat at the picture lead (±2 frames); cuts married to hits; equal spacing where the owner wants "a very clear beat" | a payoff-against-grid table; the share of cuts on onsets; off-beat payoffs listed |
|
|
207
|
+
> | 2 | Flow and transitions | every cut carries motion or a stated cause; transitions caused by the picture, never pasted; nothing randomly comes in | seam speeds in and out; dead seams; uncaused transitions |
|
|
208
|
+
> | 3 | Speed ramps | eased fast-slow-fast on every payoff and on cinematics; real slow motion, never interpolated; no constant-speed slow spans | the speed curve per payoff; the share at constant speed; the largest one-frame change |
|
|
209
|
+
> | 4 | Camera and cinematics | outside shots tied to their gameplay moment (same subject, place, frames); a clear angle; lens variety; moves into the map that hand into the primary view | per cinematic: subject, angle, tie right or wrong, every-frame motion |
|
|
210
|
+
> | 5 | Payoff moments | each payoff seen and read at a glance; its consequence completing inside the shot; framed on what animates well | frames of each payoff and reaction; reaction start within 12 frames |
|
|
211
|
+
> | 6 | Play reads as a real skilled human | movement and aim like a top player; opponents moving like players; no dead holds, searching or stuck AI; no misses | idle spans; opponent paths; misses |
|
|
212
|
+
> | 7 | HUD and text | exactly the allowed HUD, identical in every clip; text only where allowed, in the world, legible | a cross-clip HUD sheet; title crops from the finished file |
|
|
213
|
+
> | 8 | Sound mix | the music leads; payoffs level as heard; tails ring out; no transient louder than the payoffs; callouts at the owner's level and timing | a per-payoff level table; a transient ranking; the song null test |
|
|
214
|
+
> | 9 | Grade and look | one authored look that keeps the game's colours; no clipped highlights; light continuous within a sequence | luma and saturation per section; clipped share |
|
|
215
|
+
> | 10 | Pacing and structure | a shape that follows the song; clip sizes as the owner wants them (a default, a bigger one, a unique finale); no filler | payoffs per clip and section; filler shots |
|
|
216
|
+
> | 11 | Polish (nothing cheap) | no clipping, flicker, pop-in, floating parts, duplicate or stray frames on any frame | defect rows with severity |
|
|
217
|
+
> | 12 | Ending | the owner's ending: one shot, the title in the world if allowed; the song untouched | every frame of the ending; the title read at 480 px |
|
|
218
|
+
> | 13 | Overall against GOLD | would GOLD's editor sign it; would the owner share it | one paragraph; the three best and three worst moments |
|
|
219
|
+
>
|
|
220
|
+
> For every dimension add one owner-proxy line: what the owner would say, in their voice, after reading the owner-eye
|
|
221
|
+
> chapter and the ledger lines above.
|
|
222
|
+
>
|
|
223
|
+
> DELIVER {work/review/panel/{round}/{judge}.md}: (1) the thirteen scores on one line, then `OVERALL: N/10` and
|
|
224
|
+
> `Proxy score: N/10`; (2) the gap list, every dimension under 10 ranked by points lost: what is wrong (frames in OURS),
|
|
225
|
+
> what GOLD does at the equivalent moment (frames), and the smallest change that closes it, in frames, speeds, lenses and
|
|
226
|
+
> dB; (3) shot by shot: frames, what it is, a 0-10 grade, the one thing to fix; (4) every seam: clean, jerky or dead, with
|
|
227
|
+
> measured speeds; (5) must not change: what already meets the bar; (6) if your previous report is given, every earlier
|
|
228
|
+
> finding marked Fixed, Partly or Still there. End your final message with the thirteen scores on one line,
|
|
229
|
+
> `OVERALL: N/10` and `Proxy score: N/10`.
|
|
230
|
+
|
|
231
|
+
## 7. The blind critic, the owner proxy, the blind A/B and the fixer
|
|
232
|
+
|
|
233
|
+
**The blind critic** (on the plan's storyboard in the pre-mortem, and on cuts alongside the sweep):
|
|
234
|
+
|
|
235
|
+
> You are a harsh, unbiased reviewer of {an N-second gameplay montage} of {GAME, one line}, in the role of {the
|
|
236
|
+
> genre's best montage editor} and of the creative director of {the genre's official gameplay trailers}. You did not
|
|
237
|
+
> make it and have no stake in it. Be specific and severe; praise only what is at that bar. Read only: do not look at
|
|
238
|
+
> any source code, plans, notes, briefs or history; judge only the film (or the storyboard). Every note must be
|
|
239
|
+
> specific, measured where possible, and buildable.
|
|
240
|
+
>
|
|
241
|
+
> The film: {path}. The reference edit: {path} and its study {research/REF-STUDY.md}. The owner's eye: {the owner-eye
|
|
242
|
+
> chapter} and {the notes ledger}. What the owner asked for, which you must not mark down: {their standing rules,
|
|
243
|
+
> verbatim}. Notes the owner already gave on this cut, not to be reported again: {list}.
|
|
244
|
+
>
|
|
245
|
+
> Method: ffmpeg and Python only; 2 fps overviews; every-frame strips across every cut, ramp and payoff; frame-exact grabs
|
|
246
|
+
> with `select=eq(n\,N)`; flow for camera speed; beats and onsets; per-payoff in-band RMS on the mix; `nice -n 10`;
|
|
247
|
+
> never play audio aloud. Append a progress line to {log} every 20 minutes.
|
|
248
|
+
>
|
|
249
|
+
> Deliver {review/rNN/CRITIC.md}: a measured headline (payoffs on the beat at the picture lead, payoff spacing, speed
|
|
250
|
+
> curves, seam speeds, the play-to-cinematic split, hit levels as heard, the song untouched or not) with the reference's
|
|
251
|
+
> numbers beside it; scores out of 10 against the reference; at least 15 fixes with frames, what is wrong (measured), the
|
|
252
|
+
> reference moment and a buildable fix; the owner-proxy list in their voice ending in `Proxy score: N/10`; a must-not-change
|
|
253
|
+
> list; and, if a previous report exists, every earlier finding marked Fixed, Partly or Still there.
|
|
254
|
+
|
|
255
|
+
Filter every proposal through the ledger and the montage overrides before it reaches a fixer: on OG Snipers critics pushed
|
|
256
|
+
held scopes, varied gaps, designed silences and shorter tails, and the owner reversed all four.
|
|
257
|
+
|
|
258
|
+
**The owner proxy:** "Be the owner. Read the owner-eye chapter and the notes ledger in full first. For each problem: the
|
|
259
|
+
timestamp (seconds and 60 fps frame), what they would say in one blunt line, then the concrete fix. Rank by how annoyed
|
|
260
|
+
they would be. Then the things they would still ask for. End with `Proxy score: N/10`. Praise nothing they would not."
|
|
261
|
+
|
|
262
|
+
**The blind A/B of changed shots:** "Two cuts, A and B, unlabelled. Judge every changed shot and every seam: which is better
|
|
263
|
+
and why, with frames. Flag what the losing cut did better so nothing good is lost. Do not guess which is newer." Then
|
|
264
|
+
compare the whole cut against the previous one for anything lost.
|
|
265
|
+
|
|
266
|
+
**The fixer** (one per owner of a file set, per round):
|
|
267
|
+
|
|
268
|
+
> **The owner's notes, verbatim, at the top:** {quotes with ledger numbers, times, frames}. Read the common brief first.
|
|
269
|
+
> You own {slots or files}; nothing else is yours.
|
|
270
|
+
>
|
|
271
|
+
> Findings to fix (from the filtered fix list): {slot and frames}: what is wrong (measured), the target (a number), the
|
|
272
|
+
> reference moment. **First confirm each defect at the cited frames on the delivered file.** If it is not there, say so
|
|
273
|
+
> and skip it.
|
|
274
|
+
>
|
|
275
|
+
> **Do not change:** {the must-not-change list, starting with the praised list, verbatim}. Never cut a praised shot: fix
|
|
276
|
+
> it.
|
|
277
|
+
>
|
|
278
|
+
> Make the smallest change that answers the note, and sweep the class in your slots. Check on rendered frames, never by
|
|
279
|
+
> reading code. Time box: if nothing lands within about {75} minutes, finish with what exists and say so. Report in
|
|
280
|
+
> 120-300 words: what changed (frame ranges, numbers before and after), what is placeholder, what is unchecked, and the
|
|
281
|
+
> calls the owner might want to change.
|
|
282
|
+
|
|
283
|
+
## 8. Notes from the film to the game team (append-only)
|
|
284
|
+
|
|
285
|
+
Keep `handoff/GAME-NOTES.md` from day one, and ask the owner to paste its protocol into the game team's own instructions.
|
|
286
|
+
Each note has an id (F-n), the date, a priority (P1 visible in the film now and cheap-looking; P2 worked around; P3 nice
|
|
287
|
+
to have), what looks wrong, the owner's words verbatim, the evidence frames, the file and line at the frozen build, what
|
|
288
|
+
"fixed" looks like, the film's workaround and the shot it serves. The game team replies in `handoff/GAME-REPLIES.md`,
|
|
289
|
+
one line per note: the release that ships it and what changed, or "won't do, because ...". A fix the film made in its
|
|
290
|
+
own build is linked as a patch behind one flag, with before and after crops, so the game can take it as is. File a P1
|
|
291
|
+
note on day one for every payoff that depends on game behaviour (deaths, ragdolls, damage rules, opponent pathing).
|
|
292
|
+
Notes that name a cause at file and line and a measured defect ship fastest; the film never waits on the game, and
|
|
293
|
+
never edits the game team's source.
|