@milenyumai/film-kit 1.4.2 → 1.4.3

package/README.md

@@ -1,8 +1,8 @@
 # Film-Kit
 
-Professional cinematic prompt-engineering kits for Claude Code, Cursor, GitHub Copilot, and Antigravity.
+Professional cinematic prompt-engineering kits for OpenAI Codex App, Claude Code, Cursor, GitHub Copilot, and Antigravity.
 
-Film-Kit ships as a single repository with four npm packages:
+Film-Kit ships as a single repository with five npm packages:
 
 | Package | Use case | Video route |
 |---|---|---|
@@ -10,6 +10,7 @@ Film-Kit ships as a single repository with four npm packages:
 | `@milenyumai/film-kit-multi` | Multi-agent parallel production | `veo31` or `kling-3.0` |
 | `@milenyumai/film-kit-hybrid` | Nano Banana stills + Kling video | fixed `kling-3.0` |
 | `@milenyumai/film-kit-hybrid-smart` | Nano Banana stills + smart Veo/Kling routing | dialogue-aware route |
+| `@milenyumai/film-kit-studio` | Scenario to rendered media | fal.ai Nano Banana + Kling render pipeline |
 
 ## Release Highlights
 
@@ -21,16 +22,32 @@ Film-Kit ships as a single repository with four npm packages:
 - Voice-design aware audio contract with project-level `voiceCast`, shot-level `Audio Plan`, and backward-compatible `Audio direction` blocks.
 - Aligned quality gates across Claude Code, Cursor, Copilot, and Antigravity.
 - Stronger Kling 3.0 and Kling multi-shot guidance, including practical route rules and hard caps.
+- OpenAI Codex App repo-scoped surfaces across packages:
+  - `.codex/config.toml` with full-auto local profile
+  - `.codex/agents/*.toml` custom agents that point back to Film-Kit roles
+  - `.agents/skills/*/SKILL.md` Codex skill mirror
+  - AGENTS.md project-root, trusted-project, and worktree bootstrap guidance
 
 ## Editors
 
 All packages generate professional project surfaces for:
 
+- OpenAI Codex App
 - Claude Code
 - Cursor
 - GitHub Copilot
 - Antigravity
 
+## OpenAI Codex App Support
+
+Film-Kit writes repo-scoped Codex files so the Codex desktop app can discover the project from the Git root:
+
+- `.codex/config.toml` uses `approval_policy = "never"`, `sandbox_mode = "danger-full-access"`, and `[agents] max_threads = 6`, `max_depth = 1`.
+- `.codex/agents/*.toml` defines minimal custom agents that read the canonical `.agent/agents/*.md` role files.
+- `.agents/skills/*/SKILL.md` mirrors the final `.agent/skills` runtime so Codex skill discovery sees package-specific overrides.
+- Studio adds a project-scoped fal.ai MCP entry at `.codex/config.toml` with `required = false`; missing `FAL_KEY` is reported during render rather than breaking init.
+- If Codex opens a worktree without dependencies or build output, follow `AGENTS.md`: run `npm install` and `npm run build` before package commands.
+
 ## Package Guide
 
 ### 1. `@milenyumai/film-kit`
@@ -97,7 +114,8 @@ Key contracts:
 
 - one file per shot
 - route is always `kling-3.0`
-- app.kling custom storyboard cap: `3`
+- app.kling custom storyboard cap: `4`
+- Kling prompt code blocks start with `[CHARACTER / SUBJECT CONSISTENCY]`
 - hybrid-specific spatial realism rules for layered blocking
 
 ### 4. `@milenyumai/film-kit-hybrid-smart`
@@ -120,7 +138,8 @@ Key contracts:
 
 - per-section route declaration
 - Veo/Kling grammar separation
-- app.kling custom storyboard cap: `3`
+- app.kling custom storyboard cap: `4`
+- Kling-routed no-dialogue prompt code blocks start with `[CHARACTER / SUBJECT CONSISTENCY]`
 - shared spatial realism gate
 
 ## Shared Quality System
@@ -196,7 +215,7 @@ Smart hybrid:
   "defaultAspectRatio": "16:9",
   "nanoBananaImageSize": "2K",
   "klingPreset": "ultra-realism",
-  "maxKlingCustomShots": 3
+  "maxKlingCustomShots": 4
 }
 ```
 
@@ -214,6 +233,7 @@ Package publishing happens from:
 - `packages/multi` for `@milenyumai/film-kit-multi`
 - `packages/hybrid` for `@milenyumai/film-kit-hybrid`
 - `packages/hybrid-smart` for `@milenyumai/film-kit-hybrid-smart`
+- `packages/studio` for `@milenyumai/film-kit-studio`
 
 ## License
 

package/build/lib/configure.js

@@ -36,7 +36,7 @@ async function collectFiles(dir, prefix = "") {
 /**
  * Copy content files from the package's content/ directory into the target project's .agent/ directory.
  */
-async function copyContentFiles(rootDir, overwrite, templateVars) {
+async function copyContentFiles(rootDir, overwrite, templateVars, targetRoot = ".agent", skillsOnly = false) {
     const contentDir = getContentDir();
     const contentExists = await exists(contentDir);
     if (!contentExists) {
@@ -46,7 +46,10 @@ async function copyContentFiles(rootDir, overwrite, templateVars) {
     const written = [];
     const skipped = [];
     for (const file of files) {
-        const targetRelative = `.agent/${file.relativePath}`;
+        if (skillsOnly && !file.relativePath.startsWith("skills/")) {
+            continue;
+        }
+        const targetRelative = `${targetRoot}/${file.relativePath}`;
         const targetAbsolute = join(rootDir, targetRelative);
         const targetExists = await exists(targetAbsolute);
         if (targetExists && !overwrite) {
@@ -136,10 +139,19 @@ export async function configureAgents(options = {}) {
             modelDisplayName: getModelDisplayName(resolved.model)
         })
         : { written: [], skipped: [] };
+    const codexSkillMirrorResult = resolved.copyContent && resolved.platforms.includes("codex")
+        ? await copyContentFiles(resolved.rootDir, resolved.overwrite, {
+            outputDir: resolved.outputDir,
+            scenarioHint: resolved.scenarioHint,
+            model: resolved.model,
+            klingPreset: resolved.klingPreset,
+            modelDisplayName: getModelDisplayName(resolved.model)
+        }, ".agents", true)
+        : { written: [], skipped: [] };
     // 2. Generate platform-specific config files
     const files = buildProjectFiles(resolved);
-    const written = [...contentResult.written];
-    const skipped = [...contentResult.skipped];
+    const written = [...contentResult.written, ...codexSkillMirrorResult.written];
+    const skipped = [...contentResult.skipped, ...codexSkillMirrorResult.skipped];
     for (const [relativePath, content] of Object.entries(files)) {
         const absolutePath = join(resolved.rootDir, relativePath);
         const fileExists = await exists(absolutePath);

package/build/lib/defaults.js

@@ -1,4 +1,4 @@
-const ALL_PLATFORMS = ["cursor", "claude", "copilot", "antigravity"];
+const ALL_PLATFORMS = ["cursor", "claude", "copilot", "antigravity", "codex"];
 const DEFAULT_MODEL = "veo31";
 const DEFAULT_KLING_PRESET = "ultra-realism";
 export function resolveOptions(input) {

package/build/lib/templates.js

@@ -1,6 +1,25 @@
 function getModelDisplayName(model) {
     return model === "kling-3.0" ? "Kling 3.0" : "Google Flow + Veo 3.1";
 }
+function tomlString(value) {
+    return JSON.stringify(value);
+}
+function buildCodexConfig() {
+    return `approval_policy = "never"
+sandbox_mode = "danger-full-access"
+
+[agents]
+max_threads = 6
+max_depth = 1
+`;
+}
+function buildCodexAgentToml(name, description, developerInstructions) {
+    return `name = ${tomlString(name)}
+description = ${tomlString(description)}
+model_reasoning_effort = "high"
+developer_instructions = ${tomlString(developerInstructions)}
+`;
+}
 export function buildProjectFiles(config) {
     const files = {};
     // Runtime model profile file is generated from selected model options.
@@ -28,6 +47,16 @@ export function buildProjectFiles(config) {
     if (config.platforms.includes("antigravity")) {
         files[".agent/skills/shotforge-generate/SKILL.md"] = buildAntigravitySkill(config);
     }
+    if (config.platforms.includes("codex")) {
+        files[".codex/config.toml"] = buildCodexConfig();
+        files[".codex/agents/prompt-engineer.toml"] = buildCodexAgentToml("prompt_engineer", "Film-Kit prompt engineer for single-agent cinematic shot generation, repair, and delivery QA.", `Read AGENTS.md first, then .agent/model-profile.md, .agent/MASTER.md, .agent/VOICE-DESIGN.md, .agent/agents/prompt-engineer.md, and the requested .agent/workflows/*.md file before acting.
+Use .agents/skills for Codex App skill discovery; .agent remains the Film-Kit legacy runtime source.
+Keep outputs under ${config.outputDir} unless the user explicitly changes the output contract.
+For worktree runs, if node_modules or build outputs are missing and package.json exists, bootstrap with npm install and npm run build before running package commands.`);
+        if (config.platforms.includes("antigravity")) {
+            files[".agents/skills/shotforge-generate/SKILL.md"] = buildAntigravitySkill(config);
+        }
+    }
     return files;
 }
 /* ---------- AGENTS.md ---------- */
@@ -70,6 +99,14 @@ All rules, skills, and workflows are located under \`.agent/\`.
 | \`/recover\` | \`.agent/workflows/recover.md\` |
 | \`/finish\` | \`.agent/workflows/finish.md\` |
 
+## OpenAI Codex App
+- Open the Git root as the Codex project root; Codex discovers \`AGENTS.md\`, \`.codex/config.toml\`, and \`.agents/skills/\` from that root.
+- Repo-scoped Codex config lives in \`.codex/config.toml\`; Codex may ignore it until the project is trusted in the app.
+- Codex skills live under \`.agents/skills/*/SKILL.md\`; \`.agent/skills\` remains the legacy Film-Kit runtime mirror for Claude/Cursor/Copilot/Antigravity.
+- Codex custom agents live in \`.codex/agents/*.toml\` and point back to the canonical \`.agent/agents/*.md\` role files.
+- Full-auto Codex profile: \`approval_policy = "never"\`, \`sandbox_mode = "danger-full-access"\`, \`[agents] max_threads = 6\`, \`max_depth = 1\`.
+- Worktree bootstrap: if \`node_modules\` or build output is missing, run \`npm install\` and \`npm run build\` before package commands.
+
 ## Goal
 When the user asks \`/generate\`, convert the scenario into:
 - \`${config.outputDir}/project-info.md\` — Characters, settings, arc mapping
@@ -454,8 +491,8 @@ The first sentence carries the most weight. Every prompt MUST follow this order:
 - Model display: ${getModelDisplayName(config.model)}
 - Kling preset: \`${config.klingPreset}\`
 - Kling transition mode: Start+End (when model is kling-3.0)
-- Motion timeline: first → then → finally (when model is kling-3.0)
-- Kling multi-shot mode: single-transition by default; custom storyboard only for 2-3 meaningful phases (when model is kling-3.0)
+- Kling professional blocks: \`[CHARACTER / SUBJECT CONSISTENCY]\` through \`[NEGATIVE PROMPT]\` (when model is kling-3.0)
+- Kling multi-shot mode: single-transition by default; custom storyboard only for 2-4 meaningful phases, max 4 custom shots (when model is kling-3.0)
 
 ## Shot File Format (SHOTNN.md) — SINGLE FILE, ALL INCLUSIVE
 
@@ -925,34 +962,28 @@ Limit the amount of change per clip for realism:
 The official Kling app surface exposes custom storyboard-like shot prompting and optional end-frame anchoring.
 Use it as a selective upgrade, not a default:
 - **single-transition**: default mode for one clean action arc, reveal, or emotional turn
-- **custom-storyboard**: only when one clip truly needs **2-3 editorially distinct phases**
-- **hard cap in this toolkit:** 3 custom storyboard shots per video generation
+- **custom-storyboard**: only when one clip truly needs **2-4 editorially distinct phases**
+- **hard cap in this toolkit:** 4 custom storyboard shots per video generation
 - **never storyboard micro-beats:** one glance, one finger move, one tiny prop touch, one breath shift
-- if the beat needs 4+ phases, split into chained videos instead of overloading one Kling prompt
-
-### Golden Prompt Skeleton (Start+End)
-
-The prompt's job is to tell the model **how to generate the in-between frames**:
+- if the beat needs 5+ phases, split into chained videos instead of overloading one Kling prompt
 
-1. **Scene anchoring** (1 sentence)
-2. **What stays the same** (identity, costume, background stability)
-3. **Motion timeline** (2-4 steps: first → then → finally)
-4. **Camera language** (shot + movement + stabilization)
-5. **Realism constraints** (physically plausible, no warping, stable background)
-6. **Negative prompt** (artifact cleanup)
+### Golden Prompt Skeleton (Kling 3.0 Professional Multishot)
 
-**Writing "what stays the same" (critical for quality):**
-- "Same identity"
-- "Background remains stable"
-- "No rubbery motion"
-- "No melting/warping"
+Keep \`AUDIO PLAN\` and route metadata outside the prompt code block. The Kling prompt code block starts with \`[CHARACTER / SUBJECT CONSISTENCY]\` and tells the model how the scene evolves:
 
-These prevent the model from taking shortcuts.
+1. **[CHARACTER / SUBJECT CONSISTENCY]** identity, outfit, face, prop, and drift locks
+2. **[CORE INTENT]** tone and start-to-final evolution
+3. **[TOTAL DURATION]** total duration plus key action duration
+4. **[VISUAL STYLE]** time of day, lighting family, depth of field, realism level
+5. **[CAMERA]** position, angle, named 24-move lexicon movement, stabilization, and contradictions to avoid
+6. **[SHOT N | time range]** up to 4 custom storyboard shots; each shot has one job, subject action, camera behavior, background behavior, and handoff trigger
+7. **[AUDIO]** ambience, body/cloth/object sounds, main event sounds, and sound exclusions
+8. **[TIMING RULES]** real-time pacing, no slow motion/speed ramps unless intended, key action faster than the rest
+9. **[BACKGROUND RULES]** behavioral verbs for extras/environment; natural, uneven, unsynchronized motion
+10. **[FINAL FRAME PRIORITY]** last 8-12 frames converge toward the desired final composition
+11. **[NEGATIVE PROMPT]** targeted cleanup terms
 
-**Describing motion over time (model's preferred format):**
-- **First:** micro movement (breathing, blink, small head shift)
-- **Then:** main change (outfit transform / object move)
-- **Finally:** settle into end frame pose/composition
+**Timing discipline:** The total clip can be 8 seconds while the key action lasts only 0.3 seconds. Write both, so Kling does not pad fast actions with theatrical waiting.
 
 > Don't describe only the result — describe the **path** between start and end.
 
@@ -1014,12 +1045,62 @@ Use for AI artifact removal:
 
 **Template A — Minimal Realism (safest):**
 \`\`\`
-Photorealistic. Soft natural light. Same identity and stable background as the start frame. First: subtle breathing and micro facial expression changes. Then: a gentle posture shift. Finally: settle precisely into the pose and composition of the end frame. Tripod-locked shot, shallow depth of field, natural motion blur, physically plausible motion, no warping.
+[CHARACTER / SUBJECT CONSISTENCY]
+Maintain the same subject across all shots: same face, same outfit silhouette, same hair, same props. No identity drift.
+[CORE INTENT]
+A grounded cinematic transition from a quiet starting pose into a natural final pose.
+[TOTAL DURATION]
+5 seconds total; the key posture shift lasts under 1 second.
+[VISUAL STYLE]
+Photorealistic, soft natural light, shallow depth of field.
+[CAMERA]
+Tripod-locked camera, no angle jump.
+[SHOT 1 | 0.0s - 2.0s]
+Subtle breathing and micro facial expression changes; background remains stable.
+[SHOT 2 | 2.0s - 4.2s]
+The subject performs one gentle posture shift with physically plausible motion.
+[SHOT 3 | 4.2s - 5.0s]
+The subject settles into the end-frame pose and composition.
+[AUDIO]
+Natural room tone only. No music.
+[TIMING RULES]
+Real-time only. No slow motion. No speed ramps.
+[BACKGROUND RULES]
+Background remains stable and alive without jitter.
+[FINAL FRAME PRIORITY]
+Final frames match the end frame exactly.
+[NEGATIVE PROMPT]
+warping, rubbery motion, melted textures, flicker, identity drift, unwanted text, watermark
 \`\`\`
 
 **Template B — Outfit / Product Swap (commercial):**
 \`\`\`
-Photorealistic seamless swap. Same subject and pose, consistent lighting and color grade. First: slight fabric movement. Then: the outfit/product transitions smoothly with realistic material behavior (no melting, no distortion). Finally: match the end frame exactly. Stable skin texture, stable background, clean edges, no flicker.
+[CHARACTER / SUBJECT CONSISTENCY]
+Maintain the same subject, pose family, skin texture, background, and lighting. Only the outfit/product changes.
+[CORE INTENT]
+A clean commercial transformation from the start look into the end look.
+[TOTAL DURATION]
+8 seconds total; the visible swap happens smoothly over the middle 3 seconds.
+[VISUAL STYLE]
+Photorealistic commercial lighting, consistent color grade.
+[CAMERA]
+Locked camera, no orbit, no sudden angle jump.
+[SHOT 1 | 0.0s - 2.0s]
+Slight natural fabric movement; subject and background remain stable.
+[SHOT 2 | 2.0s - 5.5s]
+The outfit/product transitions with realistic material behavior, clean edges, no melting.
+[SHOT 3 | 5.5s - 8.0s]
+The subject settles into the end-frame look; lighting and scale remain consistent.
+[AUDIO]
+Subtle cloth/object texture only. No music unless requested.
+[TIMING RULES]
+Real-time only. No speed ramps.
+[BACKGROUND RULES]
+Background remains stable and secondary.
+[FINAL FRAME PRIORITY]
+Final frames match the end frame exactly.
+[NEGATIVE PROMPT]
+melting, distortion, plastic skin, flicker, identity drift, edge artifacts, text, watermark
 \`\`\`
 
 **Template C — Loop:**
@@ -1076,10 +1157,10 @@ Treat Kling multi-shot as **storyboarded internal progression**, not as a licens
 
 **Rules:**
 1. Default to **single-transition** whenever one camera path can carry the beat.
-2. Use custom storyboard only for **2-3 distinct internal phases** with different framing/action purpose.
+2. Use custom storyboard only for **2-4 distinct internal phases** with different framing/action purpose.
 3. Keep each storyboard phase short, clear, and concrete: what changed, what stayed locked, why the cut exists.
-4. Hard cap for this toolkit: **3 storyboard shots per generation** in app.kling-oriented workflows.
-5. If the sequence wants 4+ phases, split into multiple chained generations.
+4. Hard cap for this toolkit: **4 storyboard shots per generation** in app.kling-oriented workflows.
+5. If the sequence wants 5+ phases, split into multiple chained generations.
 6. Never assign separate storyboard shots to micro-actions that would read better inside one stronger shot.
 
 **Preferred storyboard jobs:**

package/build/lib/types.d.ts

@@ -1,4 +1,4 @@
-export type SupportedPlatform = "cursor" | "claude" | "copilot" | "antigravity";
+export type SupportedPlatform = "cursor" | "claude" | "copilot" | "antigravity" | "codex";
 export type SupportedModel = "veo31" | "kling-3.0";
 export type KlingPreset = "ultra-realism" | "balanced" | "custom";
 export interface AgentConfigOptions {

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

@@ -107,10 +107,10 @@ The official Kling app surface exposes start/end control, optional end-frame anc
 Use that capability with restraint:
 
 - default to a single Start+End transition prompt when the beat is one continuous motion arc, one reveal, or one emotional turn
-- use custom storyboard only when the clip truly contains **2-3 distinct editorial phases** with different visual jobs
-- in this toolkit, cap app.kling custom storyboard mode at **3 custom shots** per video
+- use custom storyboard only when the clip truly contains **2-4 distinct editorial phases** with different visual jobs
+- in this toolkit, cap app.kling custom storyboard mode at **4 custom shots** per video
 - never split one micro gesture, blink, glance, prop touch, or tiny head turn into separate storyboard shots
-- if a beat needs 4+ distinct phases, split into chained videos instead of bloating one generation
+- if a beat needs 5+ distinct phases, split into chained videos instead of bloating one generation
 - every storyboard shot must earn a clear job: establish / action / reveal / reaction / resolve
 
 ## Video Prompt Structure (Veo-Oriented Default Template)
@@ -539,20 +539,55 @@ plastic skin, waxy skin, on-screen text, watermark, logo, cartoon style, CGI loo
 > **Model-specific section.** Read `.agent/model-profile.md` for full Kling rules.
 > Skip this section if active model is Veo 3.1.
 
-### Golden Prompt Skeleton (Start+End)
-
-When using Kling Start+End mode, the prompt tells the model **how to generate in-between frames**:
-
-```
-1. [Scene anchoring] — 1 sentence: what is this scene?
-2. [What stays the same] — "Same identity", "stable background", "no rubbery motion"
-3. [Motion timeline] — "First: [micro]; Then: [main change]; Finally: [settle into end pose]"
-4. [Camera language] — shot type + movement + stabilization
-5. [Realism constraints] — "physically plausible, no warping, stable background"
-6. [Negative prompt] — artifact cleanup set
-```
-
-> **Key difference from Veo:** Don't just describe start and end — describe the **path** between them.
+### Golden Prompt Skeleton (Kling 3.0 Professional Multishot)
+
+When using Kling Start+End or custom storyboard mode, keep `AUDIO PLAN` and route metadata outside the prompt code block. The Kling prompt code block itself should start with `[CHARACTER / SUBJECT CONSISTENCY]` and read like a production brief:
+
+```
+[CHARACTER / SUBJECT CONSISTENCY]
+Maintain the same subject across all shots: same [face / silhouette / outfit / hair / accessories / identity anchors]. No identity drift, outfit drift, face drift, or prop drift.
+
+[CORE INTENT]
+A [tone] scene that evolves from [initial state] into [final state]. The video must feel [grounded / cinematic / documentary-like / tense / playful]. No [melodrama / exaggerated posing / theatrical hesitation].
+
+[TOTAL DURATION]
+[8 seconds total]. The key action lasts [0.Xs / X.Xs] and should not be stretched.
+
+[VISUAL STYLE]
+[time of day]. [lighting family]. [depth of field]. [overall realism / cinematic quality].
+
+[CAMERA]
+Use a [camera angle] camera. The camera [named 24-move lexicon movement] in a [smooth / handheld controlled / locked] manner. No contradictory camera behavior.
+
+[SHOT 1 | 0.0s - X.Xs]
+At the beginning, [starting state]. The subject [one main movement]. The camera [one behavior]. The background [behavior]. By the end, [transition trigger].
+
+[SHOT 2 | X.Xs - X.Xs]
+The shift happens. The subject [key change]. This action feels [fast / restrained / fluid / abrupt / weighty]. The background [natural unsynchronized reaction]. The camera [behavior].
+
+[SHOT 3 | X.Xs - X.Xs]
+The action continues into [next state]. The subject [movement]. The background [behavior]. This feels like [continuation / reaction / escalation], not a pose.
+
+[SHOT 4 | X.Xs - X.Xs]
+The scene resolves into [final state]. The subject [final action / stillness / expression]. The background [last state]. The camera [settle / final movement].
+
+[AUDIO]
+[ambience]. [body / cloth / object sounds]. [main event sounds]. No music, no score, no unrelated sound effects.
+
+[TIMING RULES]
+Real-time only. No slow motion. No speed ramps unless explicitly intended. The key action is faster than the rest of the clip. Only a micro-pause before the final beat.
+
+[BACKGROUND RULES]
+Background elements remain [blurred / secondary / subtle / alive]. Their behavior is [natural / uneven / unsynchronized / restrained]. No frozen extras. No repeated motion loops.
+
+[FINAL FRAME PRIORITY]
+The last [8 to 12] frames converge toward [desired final composition / reference mood / framing / emotional intensity].
+
+[NEGATIVE PROMPT]
+robotic movement, stiff animation, repeated loops, frozen extras, face drift, outfit drift, bad anatomy, frame blending, flicker, unwanted text, watermark
+```
+
+> **Key difference from Veo:** Don't just describe start and end — describe the **path** between them.
 
 ### When to Skip Prompts (Promptless Mode)
 

package/content/workflows/chain.md

@@ -42,7 +42,7 @@ All generated files must keep `SHOTNN.md` naming.
 Next shot = SHOT[last + 1]
 Chain status = CHAINED from SHOT[last]_END
 First frame code block must explicitly reuse SHOT[last]_END
-If model is kling-3.0: keep Start+End transition mode and first/then/finally motion timeline
+If model is kling-3.0: keep route metadata and `AUDIO PLAN` outside the prompt code block; the prompt code block starts with `[CHARACTER / SUBJECT CONSISTENCY]` and includes timing, background, final-frame, and negative prompt blocks
 ```
 
 ### 4. Write New Shots

package/content/workflows/generate.md

@@ -90,7 +90,7 @@ For EACH shot:
 3. If model is `kling-3.0`, perform Kling Frame Prep (see below).
 4. If model is `kling-3.0`, choose Kling execution mode:
    - `single-transition` for one strong motion arc or reveal
-   - `custom-storyboard` only for 2-3 distinct editorial phases
+   - `custom-storyboard` only for 2-4 distinct editorial phases
 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.
@@ -133,11 +133,13 @@ Before writing prompts, design the Start→End transition:
     - 5s = 1 major change
     - 10s = 2-3 staged changes
     - 15s = complex multi-step
-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 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.
+3.  **Execution mode:** Default to `single-transition`; use `custom-storyboard` only when the shot truly has 2-4 meaningful internal phases.
+4.  **Professional Kling block format:** Keep `AUDIO PLAN` and route metadata outside the prompt code block; start the Kling prompt code block with `[CHARACTER / SUBJECT CONSISTENCY]`.
+5.  **Required blocks:** Include `[CORE INTENT]`, `[TOTAL DURATION]`, `[VISUAL STYLE]`, `[CAMERA]`, `[SHOT N | time range]`, `[AUDIO]`, `[TIMING RULES]`, `[BACKGROUND RULES]`, `[FINAL FRAME PRIORITY]`, and `[NEGATIVE PROMPT]`.
+6.  **Storyboard timing:** Every custom storyboard shot must have a time range, one main job, subject action, camera behavior, background behavior, and handoff/final trigger.
+7.  **Face/hands stability:** Match orientations between start and end — avoid >45° face rotation.
+8.  **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.
+9.  **Anti-fragmentation:** Do not turn one glance, gesture, or prop touch into separate micro-shots. If custom storyboard is used, cap it at 4 stages and make each stage editorially distinct.
 
 #### Veo Gate (when model is veo31)
 
@@ -147,7 +149,8 @@ Before writing prompts, design the Start→End transition:
 
 #### Kling Gate (when model is kling-3.0)
 
-- [ ] `first → then → finally` motion timeline in VIDEO prompts
+- [ ] Kling prompt code block starts with `[CHARACTER / SUBJECT CONSISTENCY]`
+- [ ] Kling prompt includes `[TIMING RULES]`, `[BACKGROUND RULES]`, `[FINAL FRAME PRIORITY]`, and `[NEGATIVE PROMPT]`
 - [ ] "What stays the same" explicitly stated (identity, background, costume)
 - [ ] 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
@@ -156,8 +159,8 @@ Before writing prompts, design the Start→End transition:
 - [ ] Start/end frames verified for same visual universe
 - [ ] Dialogue uses `"quotation marks"` with tone markers (Kling inline format)
 - [ ] Eyelines, plane map, and shared-light logic stay consistent across start/end frames
-- [ ] Defaulted to `single-transition` unless 2-3 distinct internal phases genuinely require custom storyboard
-- [ ] If custom storyboard is used: maximum 3 stages, no decorative micro-beats, and each stage changes function/framing/action
+- [ ] Defaulted to `single-transition` unless 2-4 distinct internal phases genuinely require custom storyboard
+- [ ] If custom storyboard is used: maximum 4 stages, no decorative micro-beats, time ranges present, and each stage changes function/framing/action
 
 ### 4. Validation Pass (Mandatory Before Completion)
 

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@milenyumai/film-kit",
-  "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.",
+  "version": "1.4.3",
+  "description": "Hollywood-standard cinematic prompt engineering toolkit with model profiles (Veo 3.1 / Kling 3.0). Auto-configures AI agents (OpenAI Codex App, Cursor, Claude Code, VS Code Copilot, Antigravity) with production-grade shot generation system.",
   "type": "module",
   "main": "./build/index.js",
   "types": "./build/index.d.ts",
@@ -46,6 +46,9 @@
     "claude-code",
     "copilot",
     "antigravity",
+    "codex",
+    "openai-codex",
+    "codex-app",
     "prompt-engineering",
     "cinematic",
     "hollywood",