@kolbo/kolbo-code-linux-arm64-musl 1.1.73 → 2.0.0

package/skills/typography-video/SKILL.md

@@ -1,182 +0,0 @@
----
-name: typography-video
-description: >
-  Typography rules for video production: font selection and pairing, minimum readable sizes,
-  safe zones per platform, text animation timing, contrast requirements, subtitle styling.
-  Use when adding text overlays, titles, lower thirds, or captions to video.
-  Keywords: typography, font, text, title, lower third, caption, subtitle, safe zone, text size,
-  font pairing, readable, overlay, motion graphics
----
-
-# Typography for Video Production
-
-## Quick Reference
-
-```
-TITLE SIZE:       60-90px at 1080p  |  120-180px at 4K
-BODY SIZE:        40-60px at 1080p  |  80-120px at 4K
-SUBTITLE SIZE:    42px+ at 1080p    |  3-5% of video height
-MAX CHARS/LINE:   32-42 (subtitles) |  30 (overlays)
-MAX LINES:        2 (subtitles)     |  3 (overlays)
-READING SPEED:    21 chars/sec      |  160-200 WPM
-TITLE SAFE:       80% of frame (192px margin at 1080p)
-FONT FAMILIES:    1-2 per video maximum
-CONTRAST:         4.5:1 minimum, 7:1 optimal
-FADE DURATION:    0.3s opacity  |  0.5-1.0s slide/scale
-```
-
-## Font Selection
-
-### Recommended Video Fonts
-
-| Category | Fonts | Use For |
-|----------|-------|---------|
-| **Body / Captions** | Inter, Open Sans, Roboto, Source Sans Pro, Lato, DM Sans | All body text, subtitles |
-| **Headlines** | Montserrat Bold, Bebas Neue, Oswald Bold, Poppins Bold | Titles, section headers, key stats |
-| **Editorial** | Playfair Display, Roboto Slab | Luxury, cinematic, documentary |
-| **System Safe** | Helvetica Neue, Arial, Avenir Next | When custom fonts unavailable |
-
-### Font Pairing Rules
-
-- Limit to **1-2 font families** per video
-- Pair a **display/bold heading** font with a **neutral body** font
-- Size difference between title and body: at least **50% larger**
-- **Sans-serif** for motion graphics and captions (holds up in motion)
-- **Serif** only for cinematic title cards and editorial content
-- **Script/decorative** fonts: hero titles only, never body, never in motion
-
-### Proven Pairings
-
-| Heading | Body | Style |
-|---------|------|-------|
-| Bebas Neue | Open Sans | High-impact, social ads |
-| Montserrat Bold | Lato | Clean modern |
-| Oswald Bold | Raleway | Strong contrast |
-| Playfair Display | Inter | Editorial |
-| Poppins Bold | Poppins Light | Single-family hierarchy |
-
-## Text Sizing
-
-| Element | 1080p (px) | 4K (px) |
-|---------|-----------|---------|
-| Title / Hero text | 60-90 | 120-180 |
-| Body text | 40-60 | 80-120 |
-| Subtitles | 42+ | 84+ |
-| Lower third name | 48-60 | 96-120 |
-| Lower third role | 36-44 | 72-88 |
-
-## Safe Zones
-
-### Broadcast Standard (1920x1080)
-
-| Zone | Coverage | Margin |
-|------|----------|--------|
-| **Title Safe** | 80% of frame | 192px H, 108px V |
-| **Action Safe** | 90% of frame | 96px H, 54px V |
-
-At 1920x1080: Title Safe = inner 1536x864px
-
-### Vertical (1080x1920) Platform-Specific
-
-See the `short-form-video` skill for per-platform safe zones. Universal safe: 900x1400px centered.
-
-## Reading Speed
-
-- Average viewer reads **21 characters per second** / **160-200 WPM**
-- Minimum display time: **0.5 seconds per text element**
-- Maximum display time: **5 seconds** before it feels stale
-- Formula: `display_seconds = character_count / 21 + 0.5`
-
-## Contrast Requirements (WCAG)
-
-| Element | Minimum Ratio |
-|---------|--------------|
-| Body text on background | 4.5:1 (AA) |
-| Large text (>18pt) | 3:1 (AA) |
-| Enhanced readability | 7:1 (AAA) |
-
-**Practical rule:** After color grading, any text overlays must still meet 4.5:1 contrast against the graded background.
-
-## Text Animation Timing
-
-| Animation | Duration | Easing |
-|-----------|----------|--------|
-| Fade in/out | 0.3s | ease-in-out |
-| Slide in | 0.5-0.8s | ease-out (cubic) |
-| Scale up | 0.4-0.6s | ease-out with slight overshoot |
-| Typewriter | 30-50ms per character | linear |
-| Word-by-word | 80-120ms per word | step or ease |
-
-### Animation Rules
-- Text should be **fully readable for at least 1.5 seconds** after animation completes
-- Entry animation + readable duration + exit animation = total display time
-- **Never animate body text character-by-character** — only titles and keywords
-- **Ease out, not ease in** — text should arrive quickly and settle, not start slow
-
-## RTL (Hebrew/Arabic) Text — Proven Patterns
-
-These patterns are battle-tested in Kolbo's video production pipeline.
-
-**Reference implementation:** `./reference/KineticTitleScene.tsx` (346 lines, full RTL) — read this for a complete working example of every RTL flip listed below.
-
-### Font Selection
-- **Hebrew**: Heebo (Google Fonts, free) — use Bold weight for captions
-- **Arabic**: Cairo (Google Fonts, free) — use Bold weight for captions
-- **Never** use Poppins or English-first fonts for Hebrew/Arabic body text
-
-### Remotion RTL Rules (proven)
-
-Every component receiving text must handle RTL:
-
-```tsx
-const isHebrew = language === "he" || language === "iw";
-
-<div style={{
-  direction: isHebrew ? "rtl" : "ltr",
-  fontFamily: isHebrew ? "'Heebo', sans-serif" : "'Poppins', sans-serif",
-  textTransform: isHebrew ? "none" : "uppercase",  // No uppercase in Hebrew
-  letterSpacing: isHebrew ? 0 : -2,                 // No negative kerning for Hebrew
-}}>
-```
-
-**What you must flip for RTL:**
-| Property | LTR (English) | RTL (Hebrew) |
-|----------|--------------|--------------|
-| `direction` | `ltr` | `rtl` |
-| `paddingLeft` / `paddingRight` | normal | swapped |
-| `transformOrigin` | `"top left"` | `"top right"` |
-| Gradient direction | `90deg` | `270deg` |
-| Position "left" | left side | **right side** (flip the logic) |
-| `textTransform` | `uppercase` | `none` |
-| `letterSpacing` | `-2` or custom | `0` (always) |
-| Slide-in direction | from left | from right |
-
-### FFmpeg RTL Burn-in
-
-```bash
-# Hebrew subtitles with Heebo font
-ffmpeg -i input.mp4 -vf "subtitles=subs.srt:force_style='FontName=Heebo,FontSize=22,Bold=1,Encoding=177,PrimaryColour=&H00FFFFFF,OutlineColour=&H00000000,Outline=2'" output.mp4
-```
-
-See `subtitle-production` skill for full RTL karaoke and per-word positioning patterns.
-
----
-
-## Kolbo MCP Integration
-
-Typography rules apply to all Kolbo visual generation:
-
-| Task | Kolbo MCP Tool | Typography Rule |
-|------|---------------|----------------|
-| Title cards | `generate_image` | Add text as Remotion overlay, NOT in AI image prompt |
-| Caption generation | `transcribe_audio` | Word-level SRT → burn-in with FFmpeg |
-| Lower thirds | Remotion component | Use proven font pairings above |
-| Video with text overlay | `generate_video` + FFmpeg post | Render video first, add text in post |
-
-**Key rule:** Never ask image/video generation models to render text — they can't do it reliably. Always add text as overlays in post-production (Remotion, FFmpeg, or video editor).
-
-**Free fonts for video (Google Fonts, all free):**
-- Body: Inter, Open Sans, Roboto, Lato, DM Sans
-- Headlines: Montserrat, Bebas Neue, Oswald, Poppins
-- Hebrew: Heebo, Assistant, Rubik
-- Arabic: Cairo, Noto Sans Arabic

package/skills/typography-video/reference/KineticTitleScene.tsx

@@ -1,345 +0,0 @@
-/**
- * KineticTitleScene — Full-screen kinetic text card
- *
- * bgStyle="solid"         — always-on dark background (default)
- * bgStyle="dynamic_panel" — bg wipes in before text, wipes out after text gone
- * bgStyle="transparent"   — no background (composite over footage, render with alpha codec)
- *
- * ALL words land simultaneously (staggered by 4 frames each).
- * Alternating solid / outline / accent treatment.
- * Based on Higgsfield / SOUL 2.0 reference style.
- */
-import React from "react";
-import { useCurrentFrame, useVideoConfig, spring, interpolate, Easing } from "remotion";
-import { loadFont as loadPoppins } from "@remotion/google-fonts/Poppins";
-import { loadFont as loadHeebo } from "@remotion/google-fonts/Heebo";
-
-const { fontFamily: poppins } = loadPoppins();
-const { fontFamily: heebo } = loadHeebo();
-
-export interface KineticTitleProps {
-  words: string[];           // 2–4 words, each on its own line
-  subtext?: string;          // optional small line below
-  accentColor?: string;      // default brand blue
-  language?: string;         // "en" | "he"
-  bgStyle?: "solid" | "dynamic_panel" | "transparent";
-  durationInFrames: number;
-  fps: number;
-}
-
-const BRAND_BLUE = "#60a5fa";
-const STAGGER = 4; // frames between each word entrance
-
-// How many frames the bg leads the text on entry, and trails it on exit
-const BG_LEAD = 10;   // bg starts wiping in 10 frames before text
-const BG_TRAIL = 12;  // bg stays for 12 frames after text is gone
-
-export const KineticTitleScene: React.FC<KineticTitleProps> = ({
-  words = ["WENT", "VIRAL"],
-  subtext,
-  accentColor = BRAND_BLUE,
-  language = "en",
-  bgStyle = "solid",
-  durationInFrames = 150,
-  fps = 30,
-}) => {
-  const frame = useCurrentFrame();
-  const { width, height } = useVideoConfig();
-
-  const isHebrew = language === "he" || language === "iw";
-  const fontFamily = isHebrew ? heebo : poppins;
-  const isVertical = height > width;
-  const accent = accentColor;
-  const isDynamic = bgStyle === "dynamic_panel";
-  const isTransparent = bgStyle === "transparent";
-
-  // ── Text timing ──────────────────────────────────────────────────────────
-  // Dynamic: text starts after bg has partially appeared
-  const textOffset = isDynamic ? BG_LEAD : 0;
-  const textExitStart = durationInFrames - (isDynamic ? BG_TRAIL + 12 : 12);
-
-  const exitOpacity = interpolate(
-    frame,
-    [textExitStart, textExitStart + 12],
-    [1, 0],
-    { extrapolateRight: "clamp" }
-  );
-
-  // Global entrance scale
-  const globalScale = interpolate(frame - textOffset, [0, 20], [1.04, 1.0], {
-    extrapolateLeft: "clamp",
-    extrapolateRight: "clamp",
-    easing: Easing.out(Easing.cubic),
-  });
-
-  // ── Background panel timing ───────────────────────────────────────────────
-  // BG enters: wipes left→right starting at frame 0
-  const bgEnterDuration = 18;
-  const bgEnterProgress = interpolate(frame, [0, bgEnterDuration], [0, 1], {
-    extrapolateLeft: "clamp",
-    extrapolateRight: "clamp",
-    easing: Easing.out(Easing.cubic),
-  });
-
-  // BG exits: wipes left→right starting after text is fully gone
-  const bgExitStart = textExitStart + 12;
-  const bgExitProgress = interpolate(
-    frame,
-    [bgExitStart, bgExitStart + 14],
-    [0, 1],
-    {
-      extrapolateLeft: "clamp",
-      extrapolateRight: "clamp",
-      easing: Easing.in(Easing.cubic),
-    }
-  );
-
-  // scaleX wipe — most reliable in Remotion headless Chrome
-  // Enter: scaleX grows 0→1 from LEFT origin  (reveals left to right)
-  // Exit:  scaleX shrinks 1→0 from RIGHT origin (removes left to right)
-  const bgScaleX = isDynamic
-    ? (frame < bgExitStart
-        ? interpolate(bgEnterProgress, [0, 1], [0, 1])
-        : interpolate(bgExitProgress,  [0, 1], [1, 0]))
-    : 1;
-  const bgTransformOrigin = isDynamic && frame >= bgExitStart ? "right center" : "left center";
-
-  const showBg = !isTransparent;
-
-  // ── Accent scan line across bg (cinematic HUD feel) ──────────────────────
-  const scanY = interpolate(frame, [0, durationInFrames], [0, height], {
-    extrapolateRight: "clamp",
-  });
-  const scanOpacity = isDynamic
-    ? bgEnterProgress * (1 - bgExitProgress) * 0.12
-    : 0.06;
-
-  // ── Font size ─────────────────────────────────────────────────────────────
-  const baseFontSize = isVertical
-    ? Math.round(width * 0.22)
-    : Math.round(height * 0.26);
-
-  // ── Word styles: solid → outline → accent ────────────────────────────────
-  const wordStyles = (i: number) => {
-    const mod = i % 3;
-    if (mod === 0) return { color: "#ffffff",   stroke: "none",  shadow: `0 0 40px ${accent}44` };
-    if (mod === 1) return { color: "transparent", stroke: accent, shadow: `0 0 30px ${accent}44` };
-    return                 { color: accent,     stroke: "none",  shadow: `0 0 50px ${accent}88` };
-  };
-
-  // ── Separator line ────────────────────────────────────────────────────────
-  const SepLine = ({ delay }: { delay: number }) => {
-    const lineW = interpolate(frame - delay, [0, 16], [0, 1], {
-      extrapolateRight: "clamp",
-      easing: Easing.out(Easing.cubic),
-    });
-    return (
-      <div style={{
-        width: `${lineW * 100}%`,
-        height: 1,
-        background: `linear-gradient(${isHebrew ? "270deg" : "90deg"}, transparent, ${accent}88, transparent)`,
-        marginBottom: 2,
-        marginTop: 2,
-      }} />
-    );
-  };
-
-  return (
-    <div style={{
-      width, height,
-      // dynamic_panel: outer is transparent — the animated inner panel owns the background
-      background: (isTransparent || isDynamic) ? "transparent" : "#07070f",
-      overflow: "hidden",
-      position: "relative",
-      fontFamily,
-      direction: isHebrew ? "rtl" : "ltr",
-    }}>
-
-      {/* ── Background panel (with clip-path wipe) ── */}
-      {showBg && (
-        <div style={{
-          position: "absolute", inset: 0,
-          transform: isDynamic ? `scaleX(${bgScaleX})` : undefined,
-          transformOrigin: isDynamic ? bgTransformOrigin : undefined,
-        }}>
-          {/* Dark fill */}
-          <div style={{ position: "absolute", inset: 0, background: "#07070f" }} />
-
-          {/* Grid lines */}
-          <div style={{
-            position: "absolute", inset: 0,
-            backgroundImage: [
-              `linear-gradient(rgba(255,255,255,0.03) 1px, transparent 1px)`,
-              `linear-gradient(90deg, rgba(255,255,255,0.03) 1px, transparent 1px)`,
-            ].join(", "),
-            backgroundSize: "60px 60px",
-            pointerEvents: "none",
-          }} />
-
-          {/* Diagonal corner accent */}
-          <div style={{
-            position: "absolute",
-            top: -60,
-            left: isHebrew ? undefined : -60,
-            right: isHebrew ? -60 : undefined,
-            width: 300, height: 300,
-            background: `linear-gradient(${isHebrew ? "225deg" : "135deg"}, ${accent}18 0%, transparent 60%)`,
-            pointerEvents: "none",
-          }} />
-
-          {/* Vignette */}
-          <div style={{
-            position: "absolute", inset: 0,
-            background: "radial-gradient(ellipse at center, transparent 30%, rgba(0,0,0,0.65) 100%)",
-            pointerEvents: "none",
-          }} />
-
-          {/* Horizontal scan line (slow drift) */}
-          <div style={{
-            position: "absolute",
-            left: 0, right: 0,
-            top: scanY,
-            height: 1,
-            background: `linear-gradient(90deg, transparent, ${accent}, transparent)`,
-            opacity: scanOpacity,
-            pointerEvents: "none",
-          }} />
-
-          {/* Bright edge line at wipe front — sits at the right edge of the scaled panel */}
-          {isDynamic && bgScaleX > 0 && bgScaleX < 1 && (
-            <div style={{
-              position: "absolute",
-              top: 0, bottom: 0,
-              right: 0,
-              width: Math.round(2 / Math.max(bgScaleX, 0.05)), // compensate for scaleX compression
-              background: `linear-gradient(180deg, transparent, ${accent}, ${accent}, transparent)`,
-              opacity: 0.9,
-              boxShadow: `0 0 16px ${accent}, 0 0 32px ${accent}88`,
-            }} />
-          )}
-        </div>
-      )}
-
-      {/* ── Word stack ── */}
-      <div style={{
-        position: "absolute",
-        inset: 0,
-        display: "flex",
-        flexDirection: "column",
-        justifyContent: "center",
-        paddingLeft:  isHebrew ? 0 : isVertical ? 32 : 64,
-        paddingRight: isHebrew ? (isVertical ? 32 : 64) : 0,
-        paddingTop: isVertical ? 40 : 30,
-        transform: `scale(${globalScale})`,
-        opacity: exitOpacity,
-      }}>
-        {words.map((word, i) => {
-          const delay = textOffset + i * STAGGER;
-          const wordScale = spring({
-            frame: Math.max(0, frame - delay),
-            fps,
-            from: 0,
-            to: 1,
-            config: { damping: 9, stiffness: 350 },
-          });
-          const wordY = interpolate(
-            frame - delay, [0, 14], [40, 0],
-            { extrapolateLeft: "clamp", extrapolateRight: "clamp", easing: Easing.out(Easing.cubic) }
-          );
-          const ws = wordStyles(i);
-
-          const glitchAmt = i === 0
-            ? interpolate(frame - textOffset, [0, 1, 3], [12, 4, 0], { extrapolateRight: "clamp" })
-            : 0;
-
-          return (
-            <React.Fragment key={i}>
-              {i > 0 && <SepLine delay={textOffset + (i - 1) * STAGGER + 6} />}
-              <div style={{
-                position: "relative",
-                transform: `translateY(${wordY}px) scaleY(${wordScale})`,
-                transformOrigin: isHebrew ? "top right" : "top left",
-                lineHeight: 0.85,
-                overflow: "visible",
-              }}>
-                {glitchAmt > 0 && (
-                  <div style={{
-                    position: "absolute",
-                    fontSize: baseFontSize,
-                    fontWeight: 900,
-                    color: accent,
-                    textTransform: isHebrew ? "none" : "uppercase",
-                    letterSpacing: isHebrew ? 0 : -2,
-                    transform: `translateX(${glitchAmt}px)`,
-                    opacity: 0.45,
-                    mixBlendMode: "screen",
-                    userSelect: "none",
-                  }}>{word}</div>
-                )}
-                <div style={{
-                  fontSize: baseFontSize,
-                  fontWeight: 900,
-                  color: ws.color,
-                  textTransform: isHebrew ? "none" : "uppercase",
-                  letterSpacing: isHebrew ? 0 : -2,
-                  WebkitTextStroke: ws.stroke !== "none" ? `4px ${ws.stroke}` : undefined,
-                  textShadow: ws.shadow !== "none" ? ws.shadow : undefined,
-                  whiteSpace: "nowrap",
-                  userSelect: "none",
-                }}>{word}</div>
-              </div>
-            </React.Fragment>
-          );
-        })}
-
-        {subtext && (() => {
-          const stDelay = textOffset + words.length * STAGGER + 10;
-          const stOpacity = interpolate(frame - stDelay, [0, 14], [0, 1], {
-            extrapolateLeft: "clamp", extrapolateRight: "clamp",
-          });
-          return (
-            <div style={{
-              fontSize: Math.round(baseFontSize * 0.17),
-              fontWeight: 600,
-              color: "rgba(255,255,255,0.45)",
-              textTransform: isHebrew ? "none" : "uppercase",
-              letterSpacing: isHebrew ? 0 : 5,
-              marginTop: 20,
-              opacity: stOpacity * exitOpacity,
-            }}>{subtext}</div>
-          );
-        })()}
-      </div>
-
-      {/* ── Viewfinder corners ── */}
-      {Math.floor(frame / 10) % 2 === 0 && [
-        { top: 24, left: 24,  borderTopWidth: 2, borderLeftWidth: 2  },
-        { top: 24, right: 24, borderTopWidth: 2, borderRightWidth: 2 },
-        { bottom: 24, left: 24,  borderBottomWidth: 2, borderLeftWidth: 2  },
-        { bottom: 24, right: 24, borderBottomWidth: 2, borderRightWidth: 2 },
-      ].map((pos, i) => (
-        <div key={i} style={{
-          position: "absolute", width: 20, height: 20,
-          borderColor: accent, borderStyle: "solid", borderWidth: 0,
-          opacity: isTransparent ? exitOpacity * 0.5 : 0.5,
-          ...pos,
-        }} />
-      ))}
-
-      {/* ── Frame counter ── */}
-      <div style={{
-        position: "absolute",
-        bottom: isVertical ? 32 : 24,
-        right: isHebrew ? undefined : 36,
-        left: isHebrew ? 36 : undefined,
-        fontSize: 11,
-        fontFamily: "monospace",
-        color: `${accent}66`,
-        letterSpacing: 3,
-        opacity: exitOpacity,
-      }}>
-        {String(frame).padStart(4, "0")}
-      </div>
-    </div>
-  );
-};

package/skills/video-editing/SKILL.md

@@ -1,128 +0,0 @@
----
-name: video-editing
-description: >
-  Video editing decisions: what to cut vs keep, cut techniques (J-cut, L-cut, hard cut),
-  pacing by format, edit decision structure, silence/filler removal, talking head editing.
-  Use when making editorial decisions about video content.
-  Keywords: video editing, cut, trim, j-cut, l-cut, pacing, filler words, silence, talking head,
-  edit decision, transition, dead air, false start
----
-
-# Video Editing — Editorial Decisions
-
-## What to Cut
-
-1. **Filler words:** "um", "uh", "like", "you know" — cut at word boundaries using word timestamps
-2. **False starts:** When the speaker restarts a sentence, keep only the final take
-3. **Dead air:** Silence longer than 1.5 seconds should be trimmed to ~0.5 seconds
-4. **Off-topic tangents:** If the speaker wanders, cut to the next relevant segment
-5. **Repeated points:** Keep the best delivery, remove redundant takes
-
-## What NOT to Cut
-
-- **Breath pauses:** Natural 0.3-0.8 second pauses between sentences. These sound natural.
-- **Emphasis pauses:** Intentional pauses for dramatic effect
-- **Reactions and transitions:** Verbal bridges like "So..." or "Now..." that provide flow
-
-## Cut Techniques
-
-| Technique | Description | When to Use |
-|-----------|-------------|-------------|
-| **J-cut** | Audio from next segment starts ~0.5s before visual cut | Smooth transitions |
-| **L-cut** | Audio from current segment continues ~0.5s after visual cut | Maintaining continuity |
-| **Hard cut** | Instant transition | Major topic changes |
-| **Jump cut** | Cut within same shot (visible jump) | YouTube/social energy, pacing |
-| **Match cut** | Visual similarity bridges two different shots | Creative storytelling |
-
-## Pacing by Format
-
-| Format | Approach |
-|--------|----------|
-| **Short-form (< 60s)** | Aggressive cuts. Minimal dead air. High energy. Visual change every 1-3s |
-| **Medium-form (1-10 min)** | Balanced. Keep natural pauses for breathing room. Change every 3-5s |
-| **Long-form (> 10 min)** | Let scenes breathe. Only cut obvious problems. Change every 5-10s |
-
-## Edit Decision Structure
-
-When planning an edit, define:
-
-- **Cuts:** Ordered list of segments to keep (source, in/out points, speed)
-- **Overlays:** Timed overlay placements (images, diagrams, lower thirds)
-- **Subtitles:** Subtitle configuration (enabled, style, source file)
-- **Music:** Background music settings (asset, volume, ducking, fades)
-- **Transitions:** Transition type and timing between cuts
-
-## Silence Removal Workflow
-
-1. **Detect silence** with FFmpeg: `silencedetect=noise=-35dB:d=0.4`
-2. **Parse** silence_start/silence_end timestamps from stderr
-3. **Generate segments** of non-silent audio
-4. **Concatenate** segments with the concat demuxer
-5. Optional: apply `atempo=1.14` for subtle speedup that feels natural
-
-## Talking Head Editing Checklist
-
-- [ ] No visible jump cuts without intentional style choice
-- [ ] Audio doesn't pop or click at cut points
-- [ ] Pacing matches content energy and target platform
-- [ ] Speaker's face is never covered by overlays
-- [ ] All cuts are at word boundaries (not mid-word)
-- [ ] Filler words removed unless they serve the personality
-- [ ] B-roll covers any remaining jump cuts
-
-## Lip Sync (Dubbing / Localization)
-
-When replacing audio and matching lips:
-
-| Input | Tool | Output |
-|-------|------|--------|
-| Existing VIDEO + new audio | Lip sync | Video with synced lips |
-| Still PHOTO + audio | Talking head generator | New video from photo |
-
-**Decision rule:** If you have video footage of the person, use lip sync. If you only have a photo, use talking head generation.
-
-**Workflow for localization:**
-```
-transcribe(video) → translate → TTS(translated text) → lip_sync(original_video, new_audio)
-```
-
-Keep original video as source for each language — never chain lip sync outputs.
-
-**Face padding** for lip sync: `[0, 10, 0, 0]` (top, bottom, left, right) works for 90% of footage. Increase bottom if chin gets cropped.
-
----
-
-## Kolbo MCP Integration
-
-| Task | Kolbo MCP Tool | Notes |
-|------|---------------|-------|
-| Transcribe for edit points | `transcribe_audio` | Word-level timestamps for precise cuts |
-| Lip sync dubbing | `generate_lipsync` | Source video + new audio |
-| Generate B-roll | `generate_video` or `generate_image` | Cover jump cuts |
-| Generate narration | `generate_speech` | Re-record with AI voice |
-| Visual analysis | `chat_send_message` + Gemini | "Analyze this video for edit points" |
-
-**Editing workflow with Kolbo:**
-1. `transcribe_audio` → get full transcript with word timestamps
-2. Identify filler words, dead air, false starts from transcript
-3. Generate FFmpeg trim commands for non-silent/non-filler segments
-4. `generate_image` or `generate_video` → B-roll for covering jump cuts
-5. Concatenate clips + burn-in subtitles + mix audio
-6. Review with `production-review` skill
-
-**Localization workflow:**
-1. `transcribe_audio` → source language transcript
-2. Translate the transcript (use `chat_send_message` for translation)
-3. `generate_speech` → TTS in target language
-4. `generate_lipsync` → sync new audio to original face
-5. Repeat for each language (always from original, never chain)
-
----
-
-## Local / Free Options
-
-> **IMPORTANT:** Always use Kolbo MCP tools by default (`transcribe_audio`, `generate_lipsync`). FFmpeg silence removal is safe to use directly. For anything else, confirm with the user first.
-
-**FFmpeg (safe, standard):** Silence detection/removal, trimming, concatenation — all built-in.
-
-**Transcription:** If the user wants offline, `faster-whisper` runs on CPU (`pip install faster-whisper`). Confirm before installing.