@milenyumai/film-kit 1.4.1 → 1.4.2

package/build/lib/templates.js

@@ -48,13 +48,14 @@ All rules, skills, and workflows are located under \`.agent/\`.
 - **Model Profile:** \`.agent/model-profile.md\` — Active model rules and constraints
 - **Agent:** \`.agent/agents/prompt-engineer.md\` — Senior prompt engineer agent
 
-### Skills (8 modules)
+### Skills (9 modules)
 | Skill | Path | Priority |
 |-------|------|----------|
 | Safety Compliance | \`.agent/skills/safety-compliance/SKILL.md\` | P0 — ALWAYS |
 | Reference Locking | \`.agent/skills/reference-locking/SKILL.md\` | P1 — When refs provided |
 | Frame Chaining | \`.agent/skills/frame-chaining/SKILL.md\` | P2 — ALWAYS |
 | Spatial Blocking | \`.agent/skills/spatial-blocking/SKILL.md\` | P2 — Relational realism / gaze / depth |
+| Semantic Consistency | \`.agent/skills/semantic-consistency/SKILL.md\` | P2 — ALWAYS, visual_world + physics gate |
 | Coverage System | \`.agent/skills/coverage-system/SKILL.md\` | P2 — ALWAYS (mandatory) |
 | Visual Modes | \`.agent/skills/visual-modes/SKILL.md\` | P4 — ALWAYS |
 | Audio Design | \`.agent/skills/audio-design/SKILL.md\` | P4 — When dialogue/SFX |
@@ -75,6 +76,7 @@ When the user asks \`/generate\`, convert the scenario into:
 - \`${config.outputDir}/shot-plan.json\` — Single-agent plan + policy + \`voiceCast\` contract
 - \`${config.outputDir}/shots/SHOT01.md, SHOT02.md, ...\` — Production shot files (with coverage included)
 - \`${config.outputDir}/reports/SAFETY-REPORT.md\` — Safety gate result
+- \`${config.outputDir}/reports/SEMANTIC-REPORT.md\` — Semantic consistency gate result
 - \`${config.outputDir}/reports/DELIVERY-REPORT.md\` — Delivery gate result
 - \`${config.outputDir}/_index.md\` — Shot list with chain & status tracking
 
@@ -100,6 +102,7 @@ Each \`SHOTNN.md\` is a **single file** containing ALL shot details:
 - **Name Policy:** Visual prompts must stay anonymous. Dialogue naming follows \`shot-plan.json\` policy.
 - **AUTO-SAFETY:** Proactively reframe content that may trigger safety filters
 - **Frame Chaining:** Last frame of SHOT[N] = First frame of SHOT[N+1]
+- **Semantic Consistency:** \`shot-plan.json.visual_world\` is canonical for perspective, named camera movement strategy, shadow vector, scale, reflection, physics, and seed strategy
 - **Coverage Mandatory:** Every main shot includes 2-3 coverage sub-shots in same file
 - **Voice Design:** \`shot-plan.json\` keeps top-level \`voiceCast\`; every speaking VIDEO section keeps \`Audio Plan\`
 - **Music: NONE** by default (user must explicitly request)
@@ -122,6 +125,7 @@ Read .agent/VOICE-DESIGN.md when dialogue, narrator VO, or reusable speaker iden
 | Reference Locking | .agent/skills/reference-locking/SKILL.md | When refs provided |
 | Frame Chaining | .agent/skills/frame-chaining/SKILL.md | Multi-shot projects |
 | Spatial Blocking | .agent/skills/spatial-blocking/SKILL.md | Multi-subject / gaze / scale-critical shots |
+| Semantic Consistency | .agent/skills/semantic-consistency/SKILL.md | ALWAYS |
 | Coverage System | .agent/skills/coverage-system/SKILL.md | ALWAYS (mandatory) |
 | Visual Modes | .agent/skills/visual-modes/SKILL.md | All visual work |
 | Audio Design | .agent/skills/audio-design/SKILL.md | Dialogue/SFX needed |
@@ -137,9 +141,11 @@ Read .agent/VOICE-DESIGN.md when dialogue, narrator VO, or reusable speaker iden
 7. EVERY prompt must have an Avoid line. No exceptions.
 8. Coverage shots mandatory (2-3 per main shot, min 60 words each, included in same file).
 9. Frame chaining: Last frame of SHOT[N] = First frame of SHOT[N+1].
-10. ILK/İLK FRAME section must contain a code block even for chained shots.
-11. ONE FILE PER SHOT: Each SHOTNN.md contains main shot + all coverage shots.
-12. Keep top-level \`voiceCast\` in ${config.outputDir}/shot-plan.json and \`Audio Plan\` in every speaking VIDEO section.
+10. Semantic consistency: \`${config.outputDir}/shot-plan.json\` must include \`visual_world\`; prompts must align camera, named movement strategy, light/shadow vector, scale, reflections, physics, anatomy risk, and contextual logic.
+11. ILK/İLK FRAME section must contain a code block even for chained shots.
+12. Chained ILK/İLK FRAME code blocks must contain only: \`Use SHOT[prev]_END as exact first frame\`; any new visual prompt is a CHAIN BREAK.
+13. ONE FILE PER SHOT: Each SHOTNN.md contains main shot + all coverage shots.
+14. Keep top-level \`voiceCast\` in ${config.outputDir}/shot-plan.json and \`Audio Plan\` in every speaking VIDEO section.
 
 ## WORKFLOWS
 - /generate → Read .agent/workflows/generate.md
@@ -174,7 +180,7 @@ Read \`.agent/model-profile.md\` for active model constraints.
 
 ## SKILL LOADING (MANDATORY)
 Before generating ANY prompts:
-1. ALWAYS load: safety-compliance, frame-chaining, coverage-system, prompt-structure, visual-modes
+1. ALWAYS load: safety-compliance, frame-chaining, semantic-consistency, coverage-system, prompt-structure, visual-modes
 2. Load for relational realism: spatial-blocking
 3. Load if refs provided: reference-locking
 4. Load if dialogue/SFX: audio-design
@@ -194,9 +200,11 @@ All skills at: \`.agent/skills/[name]/SKILL.md\`
 - AUTO-ANONYMOUS: Replace ALL real names with physical descriptions
 - Dialogue naming follows \`${config.outputDir}/shot-plan.json\` policy
 - \`shot-plan.json\` stores top-level \`voiceCast\`
+- \`shot-plan.json\` stores top-level \`visual_world\` for camera/lens/camera-movement/light/shadow/scale/reflection/physics/seed strategy
 - Every speaking VIDEO section includes \`Audio Plan\`
 - AUTO-SAFETY: Proactively reframe sensitive content
 - Frame chaining: Last frame SHOT[N] = First frame SHOT[N+1]
+- Chained ILK/İLK FRAME code block contains only \`Use SHOT[prev]_END as exact first frame\`; any new visual prompt requires CHAIN BREAK
 - Coverage: 2-3 sub-shots per main shot (min 60 words each, in same file)
 - Avoid line: MANDATORY on every prompt
 - Music: NONE by default
@@ -241,6 +249,7 @@ Before generating ANY prompts, read skills from \`.agent/skills/\`:
 - \`reference-locking/SKILL.md\` — When refs provided (P1)
 - \`frame-chaining/SKILL.md\` — ALWAYS for multi-shot (P2)
 - \`spatial-blocking/SKILL.md\` — when gaze / scale / compositing realism matters (P2)
+- \`semantic-consistency/SKILL.md\` — ALWAYS, canonical \`visual_world\` + physics gate (P2)
 - \`coverage-system/SKILL.md\` — ALWAYS, mandatory (P2)
 - \`visual-modes/SKILL.md\` — ALWAYS (P4)
 - \`audio-design/SKILL.md\` — When dialogue/SFX (P4)
@@ -252,8 +261,9 @@ Before generating ANY prompts, read skills from \`.agent/skills/\`:
 - Kling preset: \`${config.klingPreset}\`
 - Create \`${config.outputDir}/project-info.md\`, \`${config.outputDir}/shot-plan.json\`, and \`${config.outputDir}/_index.md\`
 - Keep top-level \`voiceCast\` in \`${config.outputDir}/shot-plan.json\`
+- Keep top-level \`visual_world\` in \`${config.outputDir}/shot-plan.json\`
 - Write \`${config.outputDir}/shots/SHOTNN.md\` per shot; coverage stays in the same file
-- Refresh \`${config.outputDir}/reports/SAFETY-REPORT.md\` and \`${config.outputDir}/reports/DELIVERY-REPORT.md\` before \`/finish\`
+- Refresh \`${config.outputDir}/reports/SAFETY-REPORT.md\`, \`${config.outputDir}/reports/SEMANTIC-REPORT.md\`, and \`${config.outputDir}/reports/DELIVERY-REPORT.md\` before \`/finish\`
 
 ## Non-Negotiables
 1. **AUTO-ANONYMOUS:** Replace ALL real person names in visual prompts with physical descriptions.
@@ -266,10 +276,12 @@ Before generating ANY prompts, read skills from \`.agent/skills/\`:
 8. **Coverage:** 2-3 sub-shots within same SHOTNN.md file, min 70 words each.
 9. **Voice Design:** keep project-level \`voiceCast\` in \`${config.outputDir}/shot-plan.json\` and per-shot \`Audio Plan\` in each VIDEO section.
 10. **ILK/İLK FRAME:** Always include a fenced code block, even when chained.
-11. **Quality Floor:** ILK >= 80, SON >= 80, VIDEO >= 120, coverage >= 70 words.
-12. **Specificity Floor:** lens/framing, lighting, and foreground/midground/background action are mandatory.
-13. **Spatial Realism Floor:** eyeline target, plane map, shared light source, and contact/depth cues are mandatory when relational staging matters.
-14. **ONE FILE PER SHOT:** No separate coverage files.
+11. **Chained ILK/İLK FRAME:** code block contains only \`Use SHOT[prev]_END as exact first frame\`; any new visual prompt requires CHAIN BREAK.
+12. **Quality Floor:** ILK >= 80, SON >= 80, VIDEO >= 120, coverage >= 70 words.
+13. **Specificity Floor:** lens/framing, lighting, and foreground/midground/background action are mandatory.
+14. **Spatial Realism Floor:** eyeline target, plane map, shared light source, and contact/depth cues are mandatory when relational staging matters.
+15. **Semantic Consistency Floor:** \`visual_world\`, perspective/geometry, shadow vector, scale map, reflections, gravity/contact physics, anatomy risk, foreground/background coherence, contextual contradictions, and targeted semantic avoid terms are mandatory.
+16. **ONE FILE PER SHOT:** No separate coverage files.
 
 ## Workflows
 | Command | Workflow |
@@ -306,9 +318,11 @@ This workspace keeps high-level policy in \`CLAUDE.md\` and operational detail i
 - Keep one file per shot: \`${config.outputDir}/shots/SHOTNN.md\`
 - Maintain \`${config.outputDir}/shot-plan.json\` dialogue naming policy
 - Maintain \`${config.outputDir}/shot-plan.json\` top-level \`voiceCast\`
+- Maintain \`${config.outputDir}/shot-plan.json\` top-level \`visual_world\`
 - Keep \`Audio Plan\` blocks aligned to \`voiceCast\`
 - Keep \`ILK/İLK FRAME\` in a fenced code block even when chained
 - Quality floor and specificity floor are hard gates, not suggestions
+- Semantic consistency floor is a hard gate: camera/lens/camera-movement/light/shadow/scale/reflection/physics/anatomy/context must align to \`visual_world\`
 - Apply \`.agent/skills/spatial-blocking/SKILL.md\` whenever eyeline, compositing, or depth realism is critical
 
 ## Debugging
@@ -336,14 +350,17 @@ Use the Film-Kit core runtime.
 - draft and repair shot files under \`${config.outputDir}/shots/\`
 - apply \`${config.outputDir}/shot-plan.json\` dialogue naming policy
 - maintain top-level \`voiceCast\` inside \`${config.outputDir}/shot-plan.json\`
+- maintain top-level \`visual_world\` inside \`${config.outputDir}/shot-plan.json\`
 - keep \`Audio Plan\` blocks valid against \`voiceCast\`
 - enforce AUTO-ANONYMOUS, AUTO-SAFETY, chaining, and coverage contracts
 - enforce quality floor: ILK >= 80, SON >= 80, VIDEO >= 120, coverage >= 70
 - enforce specificity floor: lens/framing, lighting, and foreground/midground/background action
 - enforce spatial realism: explicit eyeline target, plane map, shared light source, and contact/depth cues when needed
+- enforce semantic consistency: \`visual_world\`, perspective/geometry, shadow vector, scale map, reflection handling, physics/anatomy risk, foreground/background coherence, contextual contradictions, and scene-specific avoid terms
 
 ## Boundaries
 - do not skip safety or delivery reports
+- do not pass chained ILK/İLK FRAME blocks that contain anything besides exact reuse text
 - do not split coverage into separate files
 - if asked to review only, report issues instead of regenerating shots by default
 `;
@@ -367,6 +384,7 @@ If using the native Claude subagent, read \`.claude/agents/prompt-engineer.md\`
    - Create \`${config.outputDir}/project-info.md\`
    - Create \`${config.outputDir}/shot-plan.json\`
    - Add top-level \`voiceCast\` before writing speaking shots
+   - Add top-level \`visual_world\` before writing visual prompts
 
 2. **Batch Strategy:**
    - 1-10 shots → Generate all at once
@@ -378,9 +396,11 @@ If using the native Claude subagent, read \`.claude/agents/prompt-engineer.md\`
    - Generate main shot (İLK FRAME + SON FRAME + VİDEO)
    - Add machine-readable \`Audio Plan\` before every VIDEO section
    - Keep İLK FRAME as fenced code block even when chained
+   - If chained, keep İLK FRAME code block to exact reuse text only; new visual prompt means CHAIN BREAK
    - Enforce hard quality floor: ILK >= 80, SON >= 80, VIDEO >= 120, coverage >= 70
    - Enforce specificity floor: lens/framing + lighting + foreground/midground/background action
    - Enforce spatial realism floor: eyeline target + plane map + shared light source + contact/depth cues when applicable
+   - Enforce semantic consistency floor: perspective/geometry + shadow vector + scale map + reflections + gravity/contact physics + anatomy risk + contextual contradiction check
    - Generate 2-3 coverage shots (in same file)
    - Write to \`${config.outputDir}/shots/SHOT[NN].md\`
    - Update \`${config.outputDir}/_index.md\`
@@ -388,6 +408,7 @@ If using the native Claude subagent, read \`.claude/agents/prompt-engineer.md\`
 4. **Validation Gates:**
    - Run /safety-check
    - Write \`${config.outputDir}/reports/SAFETY-REPORT.md\`
+   - Write \`${config.outputDir}/reports/SEMANTIC-REPORT.md\`
    - Write \`${config.outputDir}/reports/DELIVERY-REPORT.md\`
    - If any gate fails, run \`.agent/workflows/recover.md\`
 
@@ -405,12 +426,13 @@ function buildClaudeRuleOutputContract(config) {
 
 ## Required Files
 - \`${config.outputDir}/project-info.md\` — Characters, settings, emotional arc mapping, tension levels
-- \`${config.outputDir}/shot-plan.json\` — Name policy, shot plan, validation contract, and top-level \`voiceCast\`
+- \`${config.outputDir}/shot-plan.json\` — Name policy, shot plan, validation contract, top-level \`voiceCast\`, and top-level \`visual_world\`
 - \`.agent/model-profile.md\` — Active model constraints and presets
 - \`.agent/VOICE-DESIGN.md\` — Voice identity and shot audio contract
 - \`${config.outputDir}/_index.md\` — Shot tracking with chain & status
 - \`${config.outputDir}/shots/SHOT01.md ... SHOTNN.md\` — Individual shot files (one file per shot)
 - \`${config.outputDir}/reports/SAFETY-REPORT.md\` — Safety gate report
+- \`${config.outputDir}/reports/SEMANTIC-REPORT.md\` — Semantic consistency gate report
 - \`${config.outputDir}/reports/DELIVERY-REPORT.md\` — Delivery gate report
 
 ## Prompt Flow Order (MANDATORY)
@@ -449,10 +471,11 @@ FIRST SHOT / CHAINED from SHOT[prev]_END / CHAIN BREAK - Reason
 ## Main Shot
 
 ### İLK FRAME (SHOTNN_START)
-[If chained: "→ Use SHOT[prev]_END as first frame"]
+[If chained: the code block below must contain only "Use SHOT[prev]_END as exact first frame"]
 
 > NOTE: Even when chained, this section MUST contain a fenced code block.
-> If chained, include: "Use SHOT[prev]_END as exact first frame."
+> If chained, the fenced code block must contain only: "Use SHOT[prev]_END as exact first frame."
+> Any new visual prompt in a chained ILK FRAME section requires CHAIN BREAK.
 
 \\\`\\\`\\\`
 [Image prompt — min 60 words, following prompt flow order]
@@ -596,6 +619,11 @@ Character gaze directions must be spatially consistent between cuts.
 - Keep one motivated light source across subjects.
 - Add contact / weight / support cues to avoid pasted composite look.
 
+### Semantic Consistency
+- \`shot-plan.json.visual_world\` is the canonical scene contract.
+- Prompts must agree with its aspect ratio, camera height, lens family, horizon line, vanishing strategy, camera movement strategy, light source, shadow direction, color temperature, scale map, reflection risk, physics constraints, and seed strategy.
+- Avoid contextual contradictions unless the prompt explicitly explains the unusual physics or style.
+
 ### Dramaturgy (for dialogue scenes)
 Analyze per character: Objective → Obstacle → Stakes → Subtext → Beat turns.
 Embed as physical behavior in prompts, NOT as metadata.
@@ -632,10 +660,11 @@ Before generating ANY prompts, read skills from \`.agent/skills/\`:
 2. \`reference-locking/SKILL.md\` — When reference images provided
 3. \`frame-chaining/SKILL.md\` — ALWAYS for multi-shot continuity
 4. \`spatial-blocking/SKILL.md\` — When gaze / depth / scale realism is critical
-5. \`coverage-system/SKILL.md\` — ALWAYS (mandatory coverage shots)
-6. \`visual-modes/SKILL.md\` — ALWAYS (Ultra Realism default)
-7. \`audio-design/SKILL.md\` — When dialogue or SFX needed
-8. \`prompt-structure/SKILL.md\` — ALWAYS (prompt templates)
+5. \`semantic-consistency/SKILL.md\` — ALWAYS (visual_world + semantic QA)
+6. \`coverage-system/SKILL.md\` — ALWAYS (mandatory coverage shots)
+7. \`visual-modes/SKILL.md\` — ALWAYS (Ultra Realism default)
+8. \`audio-design/SKILL.md\` — When dialogue or SFX needed
+9. \`prompt-structure/SKILL.md\` — ALWAYS (prompt templates)
 
 ### When User Asks /generate
 1. Read \`.agent/workflows/generate.md\` for the full procedure
@@ -645,15 +674,18 @@ Before generating ANY prompts, read skills from \`.agent/skills/\`:
 5. Create project info: \`${config.outputDir}/project-info.md\`
 6. Create plan: \`${config.outputDir}/shot-plan.json\`
 7. Keep top-level \`voiceCast\` in the plan and \`Audio Plan\` in speaking VIDEO sections
-8. Write reports: \`${config.outputDir}/reports/SAFETY-REPORT.md\`, \`${config.outputDir}/reports/DELIVERY-REPORT.md\`
+8. Keep top-level \`visual_world\` in the plan for camera/lens/camera-movement/light/shadow/scale/reflection/physics/seed rules
+9. Write reports: \`${config.outputDir}/reports/SAFETY-REPORT.md\`, \`${config.outputDir}/reports/SEMANTIC-REPORT.md\`, \`${config.outputDir}/reports/DELIVERY-REPORT.md\`
 
 ### Critical Rules
 - **AUTO-ANONYMOUS:** Replace ALL real names with physical descriptions
 - **Name Policy:** Dialogue naming follows \`${config.outputDir}/shot-plan.json\` policy
 - **AUTO-SAFETY:** Proactively reframe sensitive content
 - **Frame Chaining:** Last frame of SHOT[N] = First frame of SHOT[N+1]
+- **Chain Hardening:** chained ILK/İLK FRAME code block contains only \`Use SHOT[prev]_END as exact first frame\`
 - **Coverage:** 2-3 sub-shots per main shot (in same file, min 60 words each)
 - **Spatial Realism:** eyeline targets, shared light, depth scale, and anti-cutout staging must agree when subjects share frame
+- **Semantic Consistency:** \`visual_world\` controls perspective/geometry, shadow vector, scale map, reflections, physics, anatomy risk, background coherence, and contextual contradictions
 - **Avoid Line:** MANDATORY on every prompt
 - **Music:** NONE by default
 - **Voice Design:** keep \`voiceCast\` in \`${config.outputDir}/shot-plan.json\` and \`Audio Plan\` in speaking VIDEO sections
@@ -686,9 +718,9 @@ When request is /generate, follow the Film-Kit Hollywood production system:
 3. Load required skills from \`.agent/skills/\`
 4. Transform scenario into production shot package at \`${config.outputDir}\`
 5. Generate: project-info.md, shot-plan.json, _index.md, shots/SHOT01.md..SHOTNN.md
-6. Keep top-level \`voiceCast\` in shot-plan.json
+6. Keep top-level \`voiceCast\` and \`visual_world\` in shot-plan.json
 7. Each SHOTNN.md: İLK FRAME + SON FRAME + AUDIO PLAN + VİDEO + 2-3 Coverage (ALL IN ONE FILE)
-8. Enforce: auto-anonymous, dialogue name policy, auto-safety, frame chaining, avoid lines
+8. Enforce: auto-anonymous, dialogue name policy, auto-safety, frame chaining, semantic consistency, avoid lines
 9. Write reports to \`${config.outputDir}/reports/\` before /finish
 `;
 }
@@ -713,10 +745,11 @@ Before generating ANY prompts, read these skills:
 2. \`.agent/skills/reference-locking/SKILL.md\` — When refs provided
 3. \`.agent/skills/frame-chaining/SKILL.md\` — ALWAYS
 4. \`.agent/skills/spatial-blocking/SKILL.md\` — When gaze / depth / scale realism is critical
-5. \`.agent/skills/coverage-system/SKILL.md\` — ALWAYS (mandatory)
-6. \`.agent/skills/visual-modes/SKILL.md\` — ALWAYS
-7. \`.agent/skills/audio-design/SKILL.md\` — When dialogue/SFX
-8. \`.agent/skills/prompt-structure/SKILL.md\` — ALWAYS
+5. \`.agent/skills/semantic-consistency/SKILL.md\` — ALWAYS (visual_world + semantic QA)
+6. \`.agent/skills/coverage-system/SKILL.md\` — ALWAYS (mandatory)
+7. \`.agent/skills/visual-modes/SKILL.md\` — ALWAYS
+8. \`.agent/skills/audio-design/SKILL.md\` — When dialogue/SFX
+9. \`.agent/skills/prompt-structure/SKILL.md\` — ALWAYS
 
 ## Workflows
 | Command | Workflow |
@@ -739,7 +772,8 @@ Before generating ANY prompts, read these skills:
 - Project info: \`${config.outputDir}/project-info.md\`
 - Plan: \`${config.outputDir}/shot-plan.json\`
 - Voice contract: top-level \`voiceCast\` in \`${config.outputDir}/shot-plan.json\`
-- Reports: \`${config.outputDir}/reports/SAFETY-REPORT.md\`, \`${config.outputDir}/reports/DELIVERY-REPORT.md\`
+- Semantic contract: top-level \`visual_world\` in \`${config.outputDir}/shot-plan.json\`
+- Reports: \`${config.outputDir}/reports/SAFETY-REPORT.md\`, \`${config.outputDir}/reports/SEMANTIC-REPORT.md\`, \`${config.outputDir}/reports/DELIVERY-REPORT.md\`
 
 ## Critical Rules
 1. **AUTO-ANONYMOUS:** Replace ALL real person names with physical descriptions
@@ -753,8 +787,10 @@ Before generating ANY prompts, read these skills:
 9. **Ultra Realism** default visual mode
 10. **8s duration** default, slow burn pacing
 11. **ILK/İLK FRAME:** always keep fenced code block
-12. **ONE FILE PER SHOT:** SHOTNN.md contains main shot + all coverage
-13. **Relational Realism:** preserve eyeline targets, shared light, depth scale, and anti-cutout staging when multiple subjects share frame
+12. **Chained ILK/İLK FRAME:** code block contains only \`Use SHOT[prev]_END as exact first frame\`; any new visual prompt is CHAIN BREAK
+13. **ONE FILE PER SHOT:** SHOTNN.md contains main shot + all coverage
+14. **Relational Realism:** preserve eyeline targets, shared light, depth scale, and anti-cutout staging when multiple subjects share frame
+15. **Semantic Consistency:** preserve \`visual_world\` perspective, shadow vector, scale map, reflections, gravity/contact physics, anatomy risk, foreground/background coherence, and contextual logic
 
 ## Quality Floor (Hard Gate)
 Reject and regenerate any shot that fails:
@@ -767,6 +803,7 @@ Reject and regenerate any shot that fails:
 - missing explicit foreground/midground/background action details
 - missing explicit eyeline target or \`not camera\` instruction when gaze matters
 - missing explicit shared light source / depth / contact cues in multi-subject shots
+- missing semantic consistency anchors: perspective/geometry, shadow vector, scale map, reflection handling, gravity/contact physics, anatomy risk, foreground/background coherence, contextual contradiction check
 
 ## Reject Weak Prompt Style
 Do not accept generic filler language:
@@ -869,7 +906,7 @@ The more aligned these are, the cleaner the transition:
 - Hand pose and finger count should be similar in both frames
 - Avoid end frames with extreme mouth positions if speech is not intended
 
-**Loop shortcut:** Set Start = End (same image). Prompt: "seamless loop" + simple camera movement (e.g., roll 360, slow push-in).
+**Loop shortcut:** Set Start = End (same image). Prompt: "seamless loop" + simple camera movement (e.g., roll 360, Dolly In).
 
 ### Transformation Budget
 
@@ -924,10 +961,11 @@ These prevent the model from taking shortcuts.
 More complex camera = more warp risk.
 
 **Safest commands (highest success rate):**
-- slow push-in / pull-back
-- pan left/right
-- tilt up/down
-- gentle handheld micro-sway
+- Dolly In / Dolly Out
+- Pan Left / Pan Right
+- Tilt Up / Tilt Down
+- Tracking Shot or Steadicam Movement for smooth follow
+- Handheld Movement with gentle micro-sway
 - roll 360 (especially for loops)
 
 **Stabilization trick:** Writing "tripod-locked" reduces background jitter.
@@ -1065,17 +1103,23 @@ Element Binding is Kling 3.0's built-in technology for maintaining character and
 - For **multi-shot** sequences: Prefer Element Binding when available
 - **Fallback:** If Element Binding is not available in your interface, manually repeat: character age, distinctive features, costume, and key proportions at each shot's start
 
-### Advanced Camera Vocabulary (Kling vCoT Triggers)
+### Advanced Camera Vocabulary (24-Move Cinematic Lexicon)
+
+These professional terms activate Kling's "Visual Chain-of-Thought" (vCoT) for more precise results. Use one named movement per shot unless a motivated compound move is required:
 
-These professional terms activate Kling's "Visual Chain-of-Thought" (vCoT) for more precise results:
+| Group | Movements |
+|-------|-----------|
+| **Physical push/pull** | Dolly In, Dolly Out |
+| **Locked head rotation** | Pan Left, Pan Right, Tilt Up, Tilt Down |
+| **Physical lateral/vertical travel** | Truck Left, Truck Right, Pedestal Up, Pedestal Down |
+| **Arc/parallax** | Arc Left, Arc Right, Tracking Shot, Leading Shot, Following Shot |
+| **Dynamic stabilization** | Whip Pan, Handheld Movement, Steadicam Movement |
+| **Angle/subjective** | Canted Angle (Dutch Angle), Point of View (POV) |
+| **Optical/composite** | Zoom In, Zoom Out, Dolly Zoom (Vertigo Effect), Crane/Jib Shot |
 
-| Category | Terms |
-|----------|-------|
-| **Angles** | Low-angle hero shot, Dutch angle (tilted horizon), POV (subjective), Bird's-eye view (top-down) |
-| **Movements** | Dolly push-in, Orbit (360° rotation), Lateral pan, Tracking, Spiral up |
-| **Hybrid** | Dolly Zoom (Vertigo effect — zoom in while pulling back), Move Left and Zoom In (simultaneous) |
+Aliases: Dolly push-in = Dolly In, Dolly pull-out = Dolly Out, Orbit = Arc Left/Right, Lateral slide = Truck Left/Right, Crane rise/descend = Crane/Jib Shot. Rack focus is a focus move, not camera travel.
 
-> **Tip:** Hybrid movements like Dolly Zoom trigger stronger vCoT processing and produce more cinematic results, but increase warp risk. Use with wider CFG (0.50-0.60).
+> **Tip:** Hybrid movements like Dolly Zoom, Truck plus Pan, or Crane/Jib plus Arc trigger stronger vCoT processing and produce more cinematic results, but increase warp risk. Use with wider CFG (0.50-0.60).
 
 ### Native Audio & Dialogue (Kling-Specific)
 

package/content/ARCHITECTURE.md

@@ -9,7 +9,7 @@
 Modular system consisting of:
 
 - **1 Specialist Agent** - Technical Prompt Engineer
-- **8 Skills** - Domain-specific knowledge modules
+- **9 Skills** - Domain-specific knowledge modules
 - **4 Workflows** - Slash command procedures
 
 ---
@@ -30,6 +30,7 @@ Modular system consisting of:
 │   ├── frame-chaining/      # Shot continuity protocol
 │   ├── coverage-system/     # Mandatory coverage shots (NEW)
 │   ├── spatial-blocking/    # Eyeline, depth, scale, compositing realism
+│   ├── semantic-consistency/ # Perspective, shadows, scale, physics, render QA
 │   ├── visual-modes/        # Ultra Realism & style modes
 │   ├── audio-design/        # Sound design rules
 │   └── prompt-structure/    # Prompt engineering patterns
@@ -47,11 +48,11 @@ Modular system consisting of:
 
 | Agent | Focus | Skills Used |
 |-------|-------|-------------|
-| `prompt-engineer` | Cinematic prompt generation for Veo 3.1 / Kling 3.0 | All 8 skills |
+| `prompt-engineer` | Cinematic prompt generation for Veo 3.1 / Kling 3.0 | All 9 skills |
 
 ---
 
-## 🧩 Skills (8)
+## 🧩 Skills (9)
 
 | Skill | Description |
 |-------|-------------|
@@ -60,6 +61,7 @@ Modular system consisting of:
 | `frame-chaining` | **Shot continuity**, last frame → first frame chaining, scene transition protocol (fade, dissolve, match cut) |
 | `coverage-system` | **Mandatory coverage shots** (Reaction, OTS, Insert, Cutaway, ECU, Wide) + L-cut/J-cut + 30° kuralı + **180° kuralı** + eyeline match + matching action + multi-character blocking |
 | `spatial-blocking` | **Relational realism**: eyeline targeting, plane mapping, body orientation, shared lighting, depth/scale integration, anti-cutout / anti-miniature cues |
+| `semantic-consistency` | **Scene-level realism gate**: canonical `visual_world`, perspective/geometry, shadow vectors, scale map, reflections, gravity/contact physics, contextual contradiction checks, render QA |
 | `visual-modes` | **Ultra Realism** default, stylization triggers, anti-AI artifact rules + **renk sürekliliği** + magic hour + flashback/rüya görsel ayrımı |
 | `audio-design` | **Sound design** rules, voice realism, project-level `voiceCast`, shot-level `audioPlan`, audio direction block + diegetic/non-diegetic ses ayrımı |
 | `prompt-structure` | Image/video prompt templates, camera vocabulary, seed parameter, prompt rewriter, **re-take strategy**, coverage prompt yazım standartları (≥60 kelime) |
@@ -91,6 +93,7 @@ User Scenario → Agent Activated → Read model-profile → Load Required Skill
                             reference-locking (if refs provided)
                             frame-chaining (ALWAYS)
                             spatial-blocking (when gaze/depth/scale realism matters)
+                            semantic-consistency (ALWAYS - visual_world + physics gate)
                             coverage-system (ALWAYS - mandatory)
                             visual-modes (check for style triggers)
                             audio-design (if dialogue/SFX needed)

package/content/MASTER.md

@@ -26,6 +26,7 @@ Scenario Received → Check for elements:
     ├── Reference images provided? → READ reference-locking/SKILL.md
     ├── Multiple shots? → READ frame-chaining/SKILL.md (ALWAYS)
     ├── Multi-subject / gaze / depth realism? → READ spatial-blocking/SKILL.md
+    ├── ALWAYS READ → semantic-consistency/SKILL.md (visual_world + semantic QA)
     ├── Style keywords (anime, noir, etc.)? → READ visual-modes/SKILL.md
     ├── Dialogue/SFX needed? → READ audio-design/SKILL.md
     ├── ALWAYS READ → coverage-system/SKILL.md (MANDATORY)
@@ -531,8 +532,9 @@ For full coverage protocols → READ `skills/coverage-system/SKILL.md`
 > 🇹🇷 [Türkçe kısa özet: Bu shot'ta ne oluyor, 1 cümle]
 
 **İLK FRAME (SHOTNN_START):**
-[If CHAINED: "→ Use SHOT[prev]_END as first frame"]
-[If FIRST/BREAK: Generate code block]
+[If CHAINED: "→ Use SHOT[prev]_END as first frame"]
+[If CHAINED: the fenced code block must contain only "Use SHOT[prev]_END as exact first frame"; any new visual prompt requires CHAIN BREAK]
+[If FIRST/BREAK: Generate code block]
 
 ```
 [Complete image prompt — MIN 60 words, MAX 100 words]
@@ -675,8 +677,9 @@ Before outputting, validate EVERY shot. **Bu kontrol otomatiktir, kullanıcı ha
 |-------|--------------| 
 | [safety-compliance](skills/safety-compliance/SKILL.md) | ALWAYS before generating |
 | [reference-locking](skills/reference-locking/SKILL.md) | When reference images provided |
-| [frame-chaining](skills/frame-chaining/SKILL.md) | ALWAYS for multi-shot |
-| [coverage-system](skills/coverage-system/SKILL.md) | ALWAYS (mandatory for every shot) |
+| [frame-chaining](skills/frame-chaining/SKILL.md) | ALWAYS for multi-shot |
+| [semantic-consistency](skills/semantic-consistency/SKILL.md) | ALWAYS for `visual_world`, perspective, shadow, scale, physics, and render QA gates |
+| [coverage-system](skills/coverage-system/SKILL.md) | ALWAYS (mandatory for every shot) |
 | [visual-modes](skills/visual-modes/SKILL.md) | Check for style triggers |
 | [audio-design](skills/audio-design/SKILL.md) | When dialogue/SFX needed |
 | [prompt-structure](skills/prompt-structure/SKILL.md) | ALWAYS |
@@ -710,18 +713,18 @@ Before outputting, validate EVERY shot. **Bu kontrol otomatiktir, kullanıcı ha
 |--------|-------------------|
 | **Composition** | Rule of thirds, golden ratio, leading lines, depth layers |
 | **Lighting** | Three-point setups, motivated sources, proper color temperature |
-| **Camera** | Professional movements (dolly, crane, Steadicam), intentional lens choices |
+| **Camera** | 24-move cinematic lexicon from `prompt-structure`, intentional lens choices |
 | **Color** | Graded for mood, consistent palette throughout |
 | **Sound** | Layered design: dialogue, SFX, ambience, Foley |
 | **Continuity** | 180° rule, eyeline match, seamless cuts |
 
 ### Professional Cinematography Terms
 
-- **Camera:** dolly, crane, Steadicam, rack focus, deep focus, tracking
+- **Camera:** Dolly In/Out, Pan Left/Right, Tilt Up/Down, Truck Left/Right, Pedestal Up/Down, Arc Left/Right, Whip Pan, Tracking Shot, Leading Shot, Following Shot, Canted Angle/Dutch Angle, Handheld Movement, Steadicam Movement, Zoom In/Out, Dolly Zoom, Crane/Jib Shot, Point of View (POV); rack focus and deep focus are focus tools, not travel moves
 - **Lighting:** key light, fill, rim/hair, practical, motivated, diffused, bounce
 - **Composition:** negative space, leading lines, frame within frame, Dutch angle
 - **Color:** LUT, color grade, desaturated, warm/cool palette, high/low key
-- **Movement:** push-in, pull-out, orbit, reveal, whip pan, crash zoom
+- **Movement:** use canonical names from `prompt-structure` instead of vague aliases; push-in = Dolly In, pull-out = Dolly Out, orbit = Arc Left/Right, crash zoom = fast Zoom In/Out
 
 ---
 

package/content/RULES.md

@@ -23,8 +23,9 @@ Film/Video request detected → Activate prompt-engineer agent
         ├── model-profile (ALWAYS FIRST)
         ├── safety-compliance (ALWAYS)
         ├── reference-locking (if refs provided)
-        ├── frame-chaining (ALWAYS for multi-shot)
-        ├── coverage-system (ALWAYS - mandatory)
+        ├── frame-chaining (ALWAYS for multi-shot)
+        ├── semantic-consistency (ALWAYS for visual_world + physics)
+        ├── coverage-system (ALWAYS - mandatory)
         ├── visual-modes (check for style triggers)
         ├── audio-design (if dialogue/SFX)
         └── prompt-structure (ALWAYS)
@@ -36,8 +37,9 @@ Film/Video request detected → Activate prompt-engineer agent
 |-------|------|--------------|
 | Safety & Celebrity Ban | `.agent/skills/safety-compliance/SKILL.md` | ALWAYS |
 | Reference Locking | `.agent/skills/reference-locking/SKILL.md` | When refs provided |
-| Frame Chaining | `.agent/skills/frame-chaining/SKILL.md` | Multi-shot projects |
-| Coverage System | `.agent/skills/coverage-system/SKILL.md` | ALWAYS (mandatory) |
+| Frame Chaining | `.agent/skills/frame-chaining/SKILL.md` | Multi-shot projects |
+| Semantic Consistency | `.agent/skills/semantic-consistency/SKILL.md` | ALWAYS |
+| Coverage System | `.agent/skills/coverage-system/SKILL.md` | ALWAYS (mandatory) |
 | Visual Modes | `.agent/skills/visual-modes/SKILL.md` | All visual work |
 | Audio Design | `.agent/skills/audio-design/SKILL.md` | Dialogue/SFX needed |
 | Prompt Structure | `.agent/skills/prompt-structure/SKILL.md` | ALWAYS |
@@ -181,7 +183,7 @@ Shot'ları zenginleştiren 12 teknik: mini hedef, fiziksel aksiyon, katmanlı ı
 |--------|-------------------|
 | **Composition** | Rule of thirds, golden ratio, leading lines, depth layers |
 | **Lighting** | Three-point, motivated sources, contrast ratios, color temperature |
-| **Camera** | Professional movements, lens selection, aperture control |
+| **Camera** | 24-move cinematic lexicon from `prompt-structure`, lens selection, aperture control |
 | **Color** | Graded for mood, consistent palette, period-appropriate |
 | **Sound** | Layered design, spatial audio, natural dynamics |
 | **Editing** | Motivated cuts, rhythm, pacing, continuity |
@@ -189,7 +191,7 @@ Shot'ları zenginleştiren 12 teknik: mini hedef, fiziksel aksiyon, katmanlı ı
 ### Professional Terms to Use
 
 ```
-Cinematography: dolly, crane, Steadicam, rack focus, deep focus
+Cinematography: Dolly In/Out, Pan Left/Right, Tilt Up/Down, Truck Left/Right, Pedestal Up/Down, Arc Left/Right, Whip Pan, Tracking Shot, Leading Shot, Following Shot, Canted Angle/Dutch Angle, Handheld Movement, Steadicam Movement, Zoom In/Out, Dolly Zoom, Crane/Jib Shot, Point of View (POV); rack focus/deep focus are focus tools
 Lighting: key light, fill, rim, practical, motivated, diffused
 Composition: negative space, leading lines, frame within frame
 Color: LUT, grade, desaturated, warm/cool, high/low key

package/content/agents/prompt-engineer.md

@@ -1,7 +1,7 @@
 ---
 name: prompt-engineer
 description: Senior Technical Prompt Engineer for model-aware runtime profiles (Veo 3.1 / Kling 3.0). Converts shot lists and scenarios into production-grade cinematic prompts optimized for continuity and platform compliance.
-skills: safety-compliance, reference-locking, frame-chaining, coverage-system, spatial-blocking, visual-modes, audio-design, prompt-structure
+skills: safety-compliance, reference-locking, frame-chaining, coverage-system, spatial-blocking, semantic-consistency, visual-modes, audio-design, prompt-structure
 ---
 
 # Prompt Engineer - Hollywood Standard Cinematic Video Generation
@@ -44,6 +44,7 @@ You are a senior Technical Prompt Engineer specialized in model-aware cinematic
 15. **Hard quality floor:** ILK >=80, SON >=80, VIDEO >=120, coverage >=70 kelime
 16. **Hard specificity floor:** Her promptta lens/framing, lighting ve FG/MG/BG action detayları zorunlu
 17. **Spatial realism floor:** eyeline target, plane map, shared light source, contact/weight cues ve tam ölçek derinlik mantığı gerektiğinde zorunlu
+18. **Semantic consistency floor:** `shot-plan.json.visual_world` kanonik olmalı; perspective/geometry, shadow vector, scale map, reflection handling, gravity/contact physics, anatomy risk, foreground/background coherence ve contextual contradiction her shot'ta geçmeli
 
 ---
 
@@ -217,12 +218,13 @@ Before outputting ANY shot:
 - [ ] Quality floor passes? (ILK>=80, SON>=80, VIDEO>=120, coverage>=70)
 - [ ] Specificity floor passes? (lens + lighting + FG/MG/BG action)
 - [ ] Spatial realism passes? (eyeline target + plane map + shared light + contact/depth cues)
+- [ ] Semantic consistency passes? (`visual_world` fields + perspective/geometry + shadow vector + scale map + reflection/physics/anatomy/context checks)
 - [ ] Model Control block exists? (`Model`, `Preset`, `CFG`, `Transition Mode`)
 
 ### 3. Kling-Specific Gates (when model is kling-3.0)
 - [ ] Motion timeline uses `first → then → finally` structure?
 - [ ] "What stays the same" explicitly stated? (identity, background, costume)
-- [ ] Camera movement is simple/safe? (no complex hybrid movements)
+- [ ] Camera movement is named from the 24-move cinematic lexicon and simple/safe unless an advanced hybrid is motivated?
 - [ ] Negative prompt includes Kling cleanup set? (warping, rubbery, melted)
 - [ ] Duration matches transformation budget? (5s=1 change, 10s=2-3, 15s=complex)
 - [ ] Start/end frames are in same visual universe? (angle, scale, light, lens)
@@ -254,11 +256,11 @@ For EACH shot, output exactly one file (`SHOTNN.md`) containing Main Shot + Cove
 
 ## Main Shot
 
-### İLK FRAME (SHOTNN_START)
-[Always provide a fenced code block]
-[If CHAINED: include "Use SHOT[prev]_END as exact first frame" inside code block]
-[If FIRST/BREAK: image prompt in same code block]
-
+### İLK FRAME (SHOTNN_START)
+[Always provide a fenced code block]
+[If CHAINED: code block must contain only "Use SHOT[prev]_END as exact first frame"; any new visual prompt requires CHAIN BREAK]
+[If FIRST/BREAK: image prompt in same code block]
+
 ### SON FRAME (SHOTNN_END)
 ```
 [Image Prompt (Flow Order + Avoid Line)]

package/content/skills/coverage-system/SKILL.md

@@ -109,7 +109,7 @@ Each coverage shot receives:
 Camera: Close-up, 85mm lens, f/2.0, shallow DOF
 Subject: Character's face, eyes, micro-expressions
 Lighting: Soft key, subtle fill, rim for separation
-Movement: Static or very slow push-in
+Movement: Static or very slow Dolly In
 Duration: 4-6 seconds
 Audio: Natural ambience only, no dialogue (listening)
 ```

package/content/skills/frame-chaining/SKILL.md

@@ -386,5 +386,5 @@ When designing END frames, consider the transformation budget:
 For seamless loops with Kling:
 1. **Start = End** (use identical image for both)
 2. Prompt: `"seamless loop"` + simple camera movement
-3. Best movements for loops: roll 360°, slow push-in/pull-back
+3. Best movements for loops: roll 360°, Dolly In/Dolly Out
 4. Avoid subject transformations — loops work best with camera-only motion

package/content/skills/prompt-structure/SKILL.md

@@ -23,7 +23,8 @@ description: Image and video prompt templates, camera movement vocabulary, stabi
 | **Reference Commands First** | Always start with reference image instructions |
 | **Safety First** | Always consider filter implications |
 | **Short Sentence Rule** | Split long sentences across shots |
-| **Model Profile First** | Read `.agent/model-profile.md` before generating any prompts |
+| **Model Profile First** | Read `.agent/model-profile.md` before generating any prompts |
+| **Semantic World First** | Read `.agent/skills/semantic-consistency/SKILL.md` and align prompts to `shot-plan.json.visual_world` |
 
 ---
 
@@ -45,9 +46,9 @@ description: Image and video prompt templates, camera movement vocabulary, stabi
 
 > **İlk cümle = "Bu shot ne?" sorusunun tek cümlelik cevabı olmalı.**
 
-### Standard Template
-
-```
+### Standard Template
+
+```
 [REFERENCE LOCK section if applicable]
 
 Cinematic still frame of [subject with reference adherence] in [frozen pose].
@@ -56,8 +57,27 @@ Lighting: [specific setup].
 Camera: [framing], [lens mm], [aperture], photorealistic, crisp focus, no motion blur, no text.
 [Safety injection if needed]
 
-[Full avoid line]
-```
+[Full avoid line]
+```
+
+### Semantic Consistency Layer
+
+Before writing any image prompt, lock the scene's `visual_world` values:
+
+- camera height, lens family, horizon line, and vanishing-point strategy
+- single motivated light source, color temperature, and shadow direction
+- foreground/midground/background scale map
+- reflection risk (`none`, `matte/non-reflective`, or accurate mirror/water/glass behavior)
+- gravity/contact physics and anatomy risk
+- contextual contradiction check
+
+For chained shots, the `ILK/İLK FRAME` fenced code block must contain only:
+
+```text
+Use SHOT[prev]_END as exact first frame.
+```
+
+Any new start-frame visual prompt must be declared as `CHAIN BREAK - [reason]`.
 
 ### Example: Character in Environment
 
@@ -146,43 +166,70 @@ Avoid: distorted faces, morphing, bad anatomy, extra limbs/fingers, blurry, flic
 
 ---
 
-## Camera Movement Vocabulary
-
-### Primary Movements
-
-| Movement | Description | Use When |
-|----------|-------------|----------|
-| **Dolly push-in** | Camera moves toward subject | Building intensity, focus |
-| **Dolly pull-out** | Camera moves away from subject | Reveal, context, ending |
-| **Pan left/right** | Camera rotates horizontally | Following action, scanning |
-| **Tilt up/down** | Camera rotates vertically | Reveal height, power dynamics |
-| **Crane rise** | Camera elevates vertically | Grand reveal, establishing |
-| **Crane descend** | Camera lowers vertically | Intimate approach |
-| **Orbit** | Camera circles subject | Dramatic emphasis, 360 view |
-| **Tracking** | Camera follows alongside | Following movement |
-| **Rack focus** | Focus shifts between planes | Attention shift, reveal |
-| **Whip pan** | Camera rotates rapidly (snap) | Transition between subjects, hide cuts, energy burst |
-| **Lateral pan** | Camera slides horizontally | Reveal adjacent elements |
-| **Spiral up** | Camera rises while orbiting | Grand cinematic reveal |
-
-### Advanced Angles (Kling vCoT Triggers)
-
-| Angle | Effect | Use When |
-|-------|--------|----------|
-| **Low-angle hero shot** | Subject towers over camera | Power, dominance, heroism |
-| **Dutch angle** | Tilted horizon | Unease, disorientation, tension |
-| **POV (subjective)** | Camera = character's eyes | Immersion, fear, discovery |
-| **Bird's-eye view** | Top-down perspective | Scale, geography, isolation |
-
-### Hybrid Movements (Advanced — Higher Risk, Higher Reward)
-
-| Movement | Effect | Risk Level |
-|----------|--------|------------|
-| **Dolly Zoom** (Vertigo) | Zoom in while pulling back | ⚠️ Medium — stunning but warp-prone |
-| **Move Left + Zoom In** | Simultaneous lateral + zoom | ⚠️ Medium — requires stable subject |
-| **Crane Rise + Pan** | Elevate while rotating | ⚠️ Low-Medium — smooth cinematic |
-
-> **Kling Note:** Professional terms activate Kling's "Visual Chain-of-Thought" (vCoT). Hybrid movements trigger stronger vCoT processing for more cinematic results, but increase warp risk. Use with CFG 0.50-0.60.
+## Camera Movement Vocabulary
+
+Use one precise movement name per shot unless the beat explicitly needs a motivated compound move. Compound moves must preserve the same `visual_world` geometry, shadow vector, screen direction, and subject scale.
+
+### Canonical 24 Cinematic Camera Movements
+
+| # | Movement | Description | Use When |
+|---|----------|-------------|----------|
+| 1 | **Dolly In** | Camera physically moves toward the subject | Build intensity, reveal emotion, enter a space |
+| 2 | **Dolly Out** | Camera physically moves away from the subject | Reveal context, isolation, ending beat |
+| 3 | **Pan Left** | Locked camera rotates left | Follow action, reveal off-screen space |
+| 4 | **Pan Right** | Locked camera rotates right | Follow action, reveal off-screen space |
+| 5 | **Tilt Up** | Locked camera rotates upward | Reveal height, power, scale |
+| 6 | **Tilt Down** | Locked camera rotates downward | Reveal ground detail, vulnerability, aftermath |
+| 7 | **Truck Left** | Whole camera slides left | Parallax, lateral reveal, follow blocking |
+| 8 | **Truck Right** | Whole camera slides right | Parallax, lateral reveal, follow blocking |
+| 9 | **Pedestal Up** | Camera physically rises vertically | Elevation reveal, character empowerment |
+| 10 | **Pedestal Down** | Camera physically lowers vertically | Intimacy, compression, descent into detail |
+| 11 | **Arc Left** | Camera moves around the subject toward screen-left | Relationship shift, dramatic parallax |
+| 12 | **Arc Right** | Camera moves around the subject toward screen-right | Relationship shift, dramatic parallax |
+| 13 | **Whip Pan** | Very fast pan with directional motion blur | Energetic transition, hidden cut, sudden discovery |
+| 14 | **Tracking Shot** | Camera follows a moving subject at matched speed | Travel, pursuit, continuous blocking |
+| 15 | **Leading Shot** | Camera moves ahead of a character, facing them | Emotional walk-and-talk, confrontation, dread |
+| 16 | **Following Shot** | Camera follows behind a character | Discovery, pursuit, subjective tension |
+| 17 | **Canted Angle (Dutch Angle)** | Camera rolls sideways, tilting the horizon | Unease, disorientation, psychological tension |
+| 18 | **Handheld Movement** | Realistic handheld shake and micro-instability | Documentary realism, urgency, chaos |
+| 19 | **Steadicam Movement** | Smooth stabilized walking movement | Long takes, premium follow shots, immersive movement |
+| 20 | **Zoom In** | Lens optically tightens without camera travel | Attention shift, surveillance, emphasis |
+| 21 | **Zoom Out** | Lens optically widens without camera travel | Reveal context, comedic or existential pullback |
+| 22 | **Dolly Zoom (Vertigo Effect)** | Camera moves while zooming the opposite way | Spatial distortion, shock, realization |
+| 23 | **Crane/Jib Shot** | Camera rises or drops on a crane/jib, often diagonally | Grand reveal, scale, transition from ground to overview |
+| 24 | **Point of View (POV)** | Camera behaves as a character's eyes | Immersion, fear, discovery, subjective action |
+
+### Legacy Aliases And Focus Moves
+
+| Existing Term | Canonical Mapping |
+|---------------|-------------------|
+| **Dolly push-in** | Dolly In |
+| **Dolly pull-out** | Dolly Out |
+| **Orbit / 360 rotation** | Arc Left / Arc Right; specify direction and avoid full 360 unless required |
+| **Lateral pan / slide** | Truck Left / Truck Right when the camera moves physically; Pan Left / Pan Right when it only rotates |
+| **Crane rise / Crane descend** | Crane/Jib Shot; use Pedestal Up/Down for a purely vertical camera lift |
+| **Spiral up** | Crane/Jib Shot + Arc Left/Right; higher warp risk |
+| **Rack focus** | Focus shift, not a camera movement; use only when foreground/background attention must change |
+
+### Advanced Angles (Kling vCoT Triggers)
+
+| Angle | Effect | Use When |
+|-------|--------|----------|
+| **Low-angle hero shot** | Subject towers over camera | Power, dominance, heroism |
+| **Dutch angle / Canted Angle** | Tilted horizon | Unease, disorientation, tension |
+| **POV / Point of View** | Camera = character's eyes | Immersion, fear, discovery |
+| **Bird's-eye view** | Top-down perspective | Scale, geography, isolation |
+
+### Hybrid Movements (Advanced - Higher Risk, Higher Reward)
+
+| Movement | Effect | Risk Level |
+|----------|--------|------------|
+| **Dolly Zoom (Vertigo Effect)** | Camera moves while the lens zooms the opposite way | Medium - stunning but warp-prone |
+| **Truck Left/Right + Pan Left/Right** | Lateral camera move plus rotation for parallax | Medium - must preserve screen direction |
+| **Crane/Jib Shot + Arc Left/Right** | Elevated diagonal move around subject | Low-Medium - smooth cinematic reveal |
+| **Whip Pan + Match Cut** | Fast blurred rotation hides transition | Medium - use only when transition is intentional |
+
+> **Kling Note:** Professional terms activate Kling's "Visual Chain-of-Thought" (vCoT). Hybrid movements trigger stronger vCoT processing for more cinematic results, but increase warp risk. Use with CFG 0.50-0.60.
 
 ---
 

package/content/skills/semantic-consistency/SKILL.md

@@ -0,0 +1,94 @@
+---
+name: semantic-consistency
+description: Enforces scene-level semantic consistency for generated image/video prompts and rendered media: perspective, shadows, scale, reflections, physics, anatomy, contextual contradictions, chain reuse, and render QA.
+---
+
+# Semantic Consistency Skill
+
+Load this skill for every Film-Kit generation, repair, safety check, and render QA pass.
+It turns semantic realism into a hard production gate instead of a best-effort prompt style.
+
+## Required `visual_world` Contract
+
+Every `shot-plan.json` or `team-plan.json` must carry a scene-level `visual_world` block.
+Use it as the canonical source for all shots in the same scene:
+
+```json
+{
+  "visual_world": {
+    "aspect_ratio": "16:9",
+    "camera_height": "eye-level / low-angle / high-angle / top-down",
+    "lens_family": "wide 24-35mm / normal 50mm / portrait 85mm / telephoto 135mm+ / macro",
+    "horizon_line": "low / center / high / not visible",
+    "vanishing_point_strategy": "single-point / two-point / flat telephoto / top-down no horizon",
+    "camera_movement_strategy": "static / one named 24-move cinematic movement / motivated compound move with parallax notes",
+    "light_source": "single motivated source, position, angle, softness",
+    "shadow_direction": "all shadows fall screen-left / screen-right / toward camera / away from camera / directly below",
+    "color_temperature": "warm 3500K / neutral daylight 5600K / cool 6500K",
+    "scale_map": "foreground, midground, background object sizes and distance logic",
+    "reflection_risk": "none / glass / mirror / water / metal, plus expected reflection behavior",
+    "physics_constraints": "gravity, contact points, support surfaces, cloth/hair/liquid behavior",
+    "seed_strategy": "locked seed for variants / new seed for structural repair / unknown platform seed"
+  }
+}
+```
+
+If any field is unknown, write an explicit value such as `not applicable: no reflective surfaces`.
+Do not leave silent blanks.
+
+## Hard Semantic Gates
+
+Reject or repair a prompt package if any item fails:
+
+- Perspective/geometry: camera angle, lens family, camera height, horizon line, vanishing logic, and camera movement strategy do not agree.
+- Camera movement: a moving shot does not use one of the 24 canonical cinematic movement names from `prompt-structure`, or combines movements without motivated parallax/scale notes.
+- Shadow vector: a single motivated light source and one consistent shadow direction are not stated.
+- Scale map: foreground/midground/background object sizes and distances are ambiguous or physically impossible.
+- Reflection handling: mirror, glass, water, or metal surfaces lack accurate reflection instructions or a `matte/non-reflective` reduction.
+- Gravity/contact physics: objects, feet, furniture, cloth, hair, liquid, smoke, and debris do not have plausible support or behavior cues.
+- Anatomy risk: visible humans lack pose simplicity, hand/face risk handling, or scene-specific anatomy avoid terms.
+- Foreground/background coherence: background does not match foreground perspective, lighting, color temperature, and style.
+- Contextual contradiction: prompt contains incompatible scene logic unless explicitly justified.
+- Scene-specific avoid line: avoid terms are generic only, missing the scene's actual failure modes.
+
+## Chained `ILK FRAME` Rule
+
+For a chained shot, the `ILK FRAME` fenced code block may contain only this reuse instruction:
+
+```text
+Use SHOT[prev]_END as exact first frame.
+```
+
+No new visual prompt, new lens, new camera angle, new lighting, or new composition is allowed inside a chained `ILK FRAME` block.
+If a new `ILK FRAME` prompt is required, mark the shot as `CHAIN BREAK - [reason]` and generate a fresh start frame.
+
+## Prompt Injection Template
+
+Append the relevant semantic anchors to every image prompt:
+
+```text
+Semantic consistency: [camera height], [lens family], [horizon/vanishing logic], camera movement [static or named 24-move lexicon term], single [light source] from [direction] at [angle], [color temperature], all shadows falling [direction], scale map [FG/MG/BG distances], contact points and gravity physically plausible, background perspective and lighting match the foreground, reflection handling [none/matte/accurate].
+Avoid: improper perspective, wrong scale, inconsistent shadows, impossible geometry, conflicting light sources, unrealistic reflections, floating objects, disconnected elements, broken gravity, bad anatomy, extra fingers, deformed hands, foreground-background mismatch, contextual contradiction, [scene-specific terms].
+```
+
+Keep avoid lines targeted. Prefer 15-25 concrete scene-relevant terms over long universal lists.
+Do not put the positive goal in the avoid line.
+
+## Render QA Gate
+
+After render, inspect the actual image/video outputs, not only the prompt text.
+Write `SEMANTIC-RENDER-REPORT.md` or a `Semantic Render QA` section in `RENDER-REPORT.md` with these fields:
+
+```markdown
+- semantic_render_status: pass/fail
+- perspective_geometry_status: pass/fail
+- shadow_vector_status: pass/fail
+- scale_depth_status: pass/fail
+- reflection_status: pass/fail/not_applicable
+- anatomy_physics_status: pass/fail
+- foreground_background_status: pass/fail
+- chain_alignment_status: pass/fail
+- rerender_or_recover_actions: [none or exact SHOTNN actions]
+```
+
+Fail the render if the actual media contradicts the canonical `visual_world` or if a chained first frame is not an exact copy of the previous rendered end frame.

package/content/skills/spatial-blocking/SKILL.md

@@ -138,6 +138,7 @@ These cues reduce the pasted-PNG feeling.
 ## 4. Shared Light Map
 
 The fastest way to create a composite artifact is to light each subject as if they live in different scenes.
+For the broader semantic gate, load `semantic-consistency/SKILL.md` and copy the canonical `visual_world.camera_movement_strategy`, `light_source`, `shadow_direction`, `color_temperature`, `scale_map`, `reflection_risk`, and `physics_constraints` into the shot language.
 
 ### Required Language
 

package/content/skills/visual-modes/SKILL.md

@@ -94,21 +94,25 @@ These artifacts identify AI-generated content. AVOID them absolutely:
 
 ## IMAGE AVOID LINE (Ultra Realism)
 
-Include at end of EVERY image prompt:
-
-```
+Include at end of EVERY image prompt:
+
+```
 Avoid: blurry, low-res, noise, jpeg artifacts, motion blur, out of focus, distorted faces, bad anatomy, extra limbs/fingers, deformed hands, mismatched eyes, warped perspective, inconsistent lighting, banding, over-sharpening, plastic skin, waxy skin, airbrushed skin, beauty filter, porcelain doll look, symmetrical face, duplicate people, floating artifacts, cutout edges, pasted composite look, toy-like scale, miniature effect, disconnected eyelines, on-screen text, captions/subtitles, watermark, logo, UI elements, cartoon/anime style, illustration style, CGI look, video game graphics, 3D render look, artificial lighting, synthetic appearance.
-```
+```
+
+Add scene-specific semantic terms from `semantic-consistency/SKILL.md` rather than blindly expanding the list. Prioritize perspective, shadow vector, scale, reflection, gravity/contact, anatomy, foreground/background, and contextual contradiction risks that actually exist in the shot.
 
 ---
 
 ## VIDEO AVOID LINE (Ultra Realism)
 
-Include at end of EVERY video prompt:
-
-```
+Include at end of EVERY video prompt:
+
+```
 Avoid: distorted faces, morphing, bad anatomy, extra limbs/fingers, blurry, flickering, frame drops, inconsistent lighting, unnatural motion, warping, rolling shutter artifacts, camera jitter, cutout edges, pasted composite look, toy-like scale, miniature effect, disconnected eyelines, on-screen text, captions/subtitles, watermark, logo, cartoon/anime style, CGI motion, synthetic appearance, robotic movement, puppet-like animation, uncanny valley expressions.
-```
+```
+
+For Start+End or image-to-video workflows, keep the semantic avoid line aligned with the same `visual_world` used by the still prompts.
 
 ---
 

package/content/workflows/chain.md

@@ -50,7 +50,9 @@ If model is kling-3.0: keep Start+End transition mode and first/then/finally mot
 - Write each shot to `$OUTPUT_DIR/shots/SHOT[NN].md`
 - Keep one-file-per-shot contract
 - Ensure `ILK/İLK FRAME` code block exists even when chained
+- For chained shots, the `ILK/İLK FRAME` code block must contain only `Use SHOT[prev]_END as exact first frame`; write `CHAIN BREAK` before any new start-frame prompt
 - Keep `Audio Plan` blocks aligned to the existing `voiceCast`
+- Keep `shot-plan.json.visual_world` consistent or document a `CHAIN BREAK` scene-world reset
 - Update `$OUTPUT_DIR/_index.md`
 
 ### 5. Refresh Reports
@@ -59,6 +61,7 @@ After continuation batch:
 
 - run `/safety-check`
 - refresh `$OUTPUT_DIR/reports/SAFETY-REPORT.md`
+- refresh `$OUTPUT_DIR/reports/SEMANTIC-REPORT.md`
 - refresh `$OUTPUT_DIR/reports/DELIVERY-REPORT.md`
 
 ---

package/content/workflows/finish.md

@@ -34,6 +34,8 @@ All shot files follow `SHOTNN.md` naming.
 7. Verify `Model Control` block exists in every shot file.
 8. Verify `shot-plan.json` has `voiceCast` coverage for every speaker or narrator.
 9. Verify every speaking VIDEO section has an `Audio Plan` block with valid `activeSpeakerKey`.
+10. Verify `shot-plan.json.visual_world` exists and every shot follows its camera, lens, camera movement strategy, shadow vector, scale map, reflection, physics, and seed strategy.
+11. Verify chained `ILK/İLK FRAME` code blocks contain only `Use SHOT[prev]_END as exact first frame`; any competing prompt must be marked `CHAIN BREAK`.
 8. For `kling-3.0`, verify:
    - `Transition Mode: Start+End`
    - CFG value is documented
@@ -44,13 +46,14 @@ All shot files follow `SHOTNN.md` naming.
 Required reports:
 
 - `$OUTPUT_DIR/reports/SAFETY-REPORT.md`
+- `$OUTPUT_DIR/reports/SEMANTIC-REPORT.md`
 - `$OUTPUT_DIR/reports/DELIVERY-REPORT.md`
 
 Gate rules:
 
 - reports must exist
 - reports must have non-contradictory status fields
-- both reports must be pass
+- all reports must be pass
 
 If any rule fails, run `/recover` and do not finish.
 
@@ -82,6 +85,7 @@ Do not declare completion unless:
 
 - all shot files pass structure and continuity
 - safety report is pass
+- semantic report is pass
 - delivery report is pass
 - final summary says pass
 - model-specific checks pass (Kling or Veo profile)

package/content/workflows/generate.md

@@ -38,6 +38,7 @@ $OUTPUT_DIR/
 │   └── ...
 ├── reports/
 │   ├── SAFETY-REPORT.md     # Safety and policy validation
+│   ├── SEMANTIC-REPORT.md   # Perspective, shadow, scale, physics validation
 │   └── DELIVERY-REPORT.md   # Final packaging validation
 └── _index.md                # Shot list with status and report gate table
 ```
@@ -64,6 +65,7 @@ $OUTPUT_DIR/
    - `dialogue_name_policy` (`preserve-original-dialogue` or `anonymize-dialogue`)
    - top-level `voiceCast`
    - voice defaults (`single_active_speaker`, `music_default`, `subtitles_default`)
+   - top-level `visual_world` with `aspect_ratio`, `camera_height`, `lens_family`, `horizon_line`, `vanishing_point_strategy`, `camera_movement_strategy`, `light_source`, `shadow_direction`, `color_temperature`, `scale_map`, `reflection_risk`, `physics_constraints`, and `seed_strategy`
 10. Ensure every speaker or narrator has a stable `speakerKey` before shot writing.
 11. Ensure directories exist: `$OUTPUT_DIR/`, `$OUTPUT_DIR/shots/`, `$OUTPUT_DIR/reports/`.
 
@@ -92,7 +94,7 @@ For EACH shot:
 4. Reuse or create the correct `voiceCast` entry for any speaking character or narrator.
 5. Generate main shot prompts (`ILK/İLK FRAME`, `SON FRAME`, `VIDEO`).
 5. `ILK/İLK FRAME` section MUST always include a fenced code block.
-6. If chained, first-frame code block must explicitly state: `Use SHOT[prev]_END as exact first frame`.
+6. If chained, first-frame code block must contain only: `Use SHOT[prev]_END as exact first frame`. Any new visual prompt, camera, lens, lighting, or composition requires `CHAIN BREAK - [reason]`.
 7. Write a machine-readable `Audio Plan` JSON block for every VIDEO section.
 8. If dialogue or voiceover exists, require `activeSpeakerKey`, `dialogueLines`, and `performanceNote`.
 9. Keep one active speaker per shot. Split reply dialogue across multiple shots.
@@ -103,11 +105,19 @@ For EACH shot:
    - each coverage prompt: minimum 70 words
 8. Enforce specificity floor on every prompt:
    - explicit lens/framing/camera movement
+   - camera movement selected from the 24-move cinematic lexicon in `prompt-structure` when movement is present
    - explicit lighting direction/intensity/atmosphere
    - explicit foreground/midground/background action details
    - explicit eyeline target / body orientation when a subject looks at someone or something
    - explicit shared light source / bounce logic when multiple subjects share frame
    - explicit depth/scale integration when more than one plane is visible
+9. Enforce semantic consistency floor on every prompt:
+   - perspective/geometry matches `shot-plan.json -> visual_world`
+   - shadow vector follows the canonical light source and `shadow_direction`
+   - scale map and foreground/midground/background distances are physically plausible
+   - reflection handling is accurate or reflective surfaces are intentionally avoided
+   - gravity/contact physics, anatomy risk, foreground/background coherence, and contextual contradictions are resolved
+   - avoid line contains targeted scene-specific semantic failure terms
 10. Generate coverage prompts (2-3 per main shot, min 70 words each).
 11. Add Turkish summary for shot and each coverage section.
 12. Apply model-specific generation gates (see below).
@@ -126,7 +136,7 @@ Before writing prompts, design the Start→End transition:
 3.  **Execution mode:** Default to `single-transition`; use `custom-storyboard` only when the shot truly has 2-3 meaningful internal phases.
 4.  **Motion timeline:** Write 2-4 steps: `first → then → finally`.
 5.  **Face/hands stability:** Match orientations between start and end — avoid >45° face rotation.
-6.  **Camera safety:** Use only safe movements (slow push-in, pan, tilt, micro-sway, tripod-locked).
+6.  **Camera safety:** Use the 24-move cinematic lexicon; prefer safe simple moves (Dolly In/Out, Pan Left/Right, Tilt Up/Down, Truck Left/Right, Pedestal Up/Down, Tracking Shot, Steadicam Movement) before advanced hybrids.
 7.  **Anti-fragmentation:** Do not turn one glance, gesture, or prop touch into separate micro-shots. If custom storyboard is used, cap it at 3 stages and make each stage editorially distinct.
 
 #### Veo Gate (when model is veo31)
@@ -139,7 +149,7 @@ Before writing prompts, design the Start→End transition:
 
 - [ ] `first → then → finally` motion timeline in VIDEO prompts
 - [ ] "What stays the same" explicitly stated (identity, background, costume)
-- [ ] Camera movement is simple and safe (no complex hybrid)
+- [ ] Camera movement is named from the 24-move cinematic lexicon and kept simple/safe unless the beat requires an advanced hybrid
 - [ ] `stable background`, `no warping`, `physically plausible` constraints
 - [ ] Kling negative prompt set active (warping, rubbery, melted, deformed)
 - [ ] Duration matches transformation budget
@@ -162,8 +172,18 @@ Before writing prompts, design the Start→End transition:
    - `file_completeness_status`
    - `packaging_status`
    - `blockers`
-4. Reject any shot output that passes while violating quality floor or specificity floor.
-5. Do not finalize if any report is fail or missing.
+4. Write `$OUTPUT_DIR/reports/SEMANTIC-REPORT.md` with strict fields:
+   - `overall_status`
+   - `visual_world_status`
+   - `perspective_geometry_status`
+   - `shadow_vector_status`
+   - `scale_depth_status`
+   - `reflection_physics_anatomy_status`
+   - `contextual_contradiction_status`
+   - `chained_ilk_frame_status`
+   - `blockers`
+5. Reject any shot output that passes while violating quality floor, specificity floor, or semantic consistency floor.
+6. Do not finalize if any report is fail or missing.
 
 ---
 
@@ -192,8 +212,8 @@ FIRST SHOT / CHAINED from SHOT[prev]_END / CHAIN BREAK - Reason
 ### ILK FRAME (SHOTNN_START)
 
 ```text
-[Image prompt - min 80 words, Flow Order]
-[If chained: include "Use SHOT[prev]_END as exact first frame"]
+[Image prompt - min 80 words, Flow Order for FIRST SHOT or CHAIN BREAK only]
+[If chained: the entire code block must be only "Use SHOT[prev]_END as exact first frame"]
 Avoid: blurry, low-res, text, watermark, bad anatomy, distorted face ...
 ```
 
@@ -305,6 +325,7 @@ Report:
 [N] shot olusturuldu ve $OUTPUT_DIR/shots/ klasorune kaydedildi.
 Toplam: [N] ana shot + [M] coverage = [N+M] production shots.
 Safety report: $OUTPUT_DIR/reports/SAFETY-REPORT.md
+Semantic report: $OUTPUT_DIR/reports/SEMANTIC-REPORT.md
 Delivery report: $OUTPUT_DIR/reports/DELIVERY-REPORT.md
 
 Devam etmek icin: 'devam et' veya '/chain'

package/content/workflows/recover.md

@@ -12,9 +12,12 @@ $ARGUMENTS
 
 - safety report is fail
 - delivery report is fail
+- semantic report is fail
 - missing required shot sections
 - continuity mismatch between neighboring shots
 - missing `voiceCast` entry or broken `activeSpeakerKey` binding
+- missing or contradictory `shot-plan.json.visual_world`
+- chained `ILK/İLK FRAME` block contains a new visual prompt instead of exact reuse
 
 ## Recovery Steps
 
@@ -24,11 +27,13 @@ $ARGUMENTS
 4. Repair `voiceCast` or `Audio Plan` bindings before rerunning reports.
 5. Re-run `/safety-check`.
 6. Regenerate `$OUTPUT_DIR/reports/DELIVERY-REPORT.md`.
-7. Update `$OUTPUT_DIR/_index.md` with recovered status.
+7. Regenerate `$OUTPUT_DIR/reports/SEMANTIC-REPORT.md`.
+8. Update `$OUTPUT_DIR/_index.md` with recovered status.
 
 ## Exit Criteria
 
 Recovery is complete only when both are pass:
 
 - `$OUTPUT_DIR/reports/SAFETY-REPORT.md`
+- `$OUTPUT_DIR/reports/SEMANTIC-REPORT.md`
 - `$OUTPUT_DIR/reports/DELIVERY-REPORT.md`

package/content/workflows/safety-check.md

@@ -46,8 +46,18 @@ Validate all prompts before delivery to ensure platform compliance.
 3. Continuity checks
 - `SHOT[N]_END` aligns with `SHOT[N+1]_START`
 - chain breaks are explicitly declared when needed
-
-4. Quality checks
+- chained `ILK/İLK FRAME` code blocks contain only `Use SHOT[prev]_END as exact first frame`; competing new prompts are automatic fail
+
+4. Semantic consistency checks
+- `shot-plan.json` contains top-level `visual_world`
+- `visual_world` includes `aspect_ratio`, `camera_height`, `lens_family`, `horizon_line`, `vanishing_point_strategy`, `camera_movement_strategy`, `light_source`, `shadow_direction`, `color_temperature`, `scale_map`, `reflection_risk`, `physics_constraints`, and `seed_strategy`
+- every prompt aligns perspective/geometry with the canonical lens/camera/horizon/vanishing logic
+- every moving prompt uses one of the 24 canonical cinematic camera movement names from `prompt-structure`
+- every prompt states one shadow vector from the canonical light source
+- scale/depth, foreground/background coherence, reflection handling, gravity/contact physics, anatomy risk, and contextual contradictions are resolved
+- avoid lines include targeted scene-specific semantic failure terms
+
+5. Quality checks
 - `ILK/İLK FRAME` prompt >= 80 words
 - `SON FRAME` prompt >= 80 words
 - `VIDEO/VİDEO` prompt >= 120 words
@@ -60,7 +70,7 @@ Validate all prompts before delivery to ensure platform compliance.
 - contact / weight / support cues exist when compositing realism is critical
 - depth / scale integration is explicit when multiple planes are visible
 
-5. Kling-only checks (when model is `kling-3.0`)
+6. Kling-only checks (when model is `kling-3.0`)
 - `Model Control` block exists in each SHOT file
 - `Transition Mode: Start+End` is declared
 - VIDEO prompt contains explicit `first -> then -> finally` progression
@@ -90,6 +100,30 @@ Write `$OUTPUT_DIR/reports/SAFETY-REPORT.md` using strict fields:
 - [concise bullet list or none]
 ```
 
+Write `$OUTPUT_DIR/reports/SEMANTIC-REPORT.md` using strict fields:
+
+```markdown
+# SEMANTIC REPORT
+
+- overall_status: pass/fail
+- visual_world_status: pass/fail
+- perspective_geometry_status: pass/fail
+- shadow_vector_status: pass/fail
+- scale_depth_status: pass/fail
+- reflection_physics_anatomy_status: pass/fail
+- foreground_background_status: pass/fail
+- contextual_contradiction_status: pass/fail
+- chained_ilk_frame_status: pass/fail
+- blockers:
+  - none
+
+## Findings
+- [concise bullet list]
+
+## Fixes Applied
+- [concise bullet list or none]
+```
+
 Rules:
 - Status fields must not contradict each other.
 - If any status is fail, `overall_status` must be fail.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@milenyumai/film-kit",
-  "version": "1.4.1",
+  "version": "1.4.2",
   "description": "Hollywood-standard cinematic prompt engineering toolkit with model profiles (Veo 3.1 / Kling 3.0). Auto-configures AI agents (Cursor, Claude Code, VS Code Copilot, Antigravity) with production-grade shot generation system.",
   "type": "module",
   "main": "./build/index.js",