Skip to content

video-use Best Practices (Hebrew)

Trusted88/100
Before deciding whether to install, talk to the skill

Hebrew override layer on top of browser-use/video-use. Two paths: (1) captions-only.sh, one command that takes a long video, transcribes via Scribe, and burns Hebrew captions on the original. Costs $1-3 regardless of video length, the cheap path for non-technical users with a webinar or lecture. (2) The full video-use cut workflow for editing raw footage into curated teasers, ~$25-300 depending on length. Also covers SUB_FORCE_STYLE override (Helvetica has no Hebrew glyphs), python-bidi pre-shape recipe (libass+SRT BiDi is silently broken on macOS), Scribe garbage-character sanitization (Devanagari and other non-Hebrew chars Scribe mixes into Hebrew transcripts), Hebrew filler-word lexicon, explicit fontsdir= for reliable font discovery, and frame-sampling self-eval protocol. Use for any Hebrew video that needs captions or editing with video-use. Do NOT use for non-Hebrew video-use sessions, Hebrew podcast audio-only post-production (use hebrew-podcast-postproduction), or generic FFmpeg work without video-use.

Trust score 88/100 (Trusted) · 10+ installs · 3 GitHub contributors · MIT license

The Problem

video-use ships with a strong English-first default: the bundled SUB_FORCE_STYLE uses Helvetica (no Hebrew glyphs, renders as boxes), 2-word UPPERCASE chunks (Hebrew has no uppercase), and ElevenLabs Scribe filler tagging is English-centric. This skill is the Hebrew-specific override layer on top of video-use 12 Hard Rules, none of the rules change, only the style values, font fallback chain, and filler list.

skills-ilskills-ilDeveloper Tools
1.2.9MITGitHub
10installs582views
0Write a Review
npx skills-il add skills-il/developer-tools@v1.2.9-video-use-best-practices --skill video-use-best-practices -a claude-code
Install on Claude.ai, Claude Desktop, ChatGPT, Manus, or other platforms

This skill doesn't support Claude Desktop. It works with CLI-based agents only.

A new version released? How to update your installed skill
Not sure how? Read the guide

When to Apply

  • When editing a Hebrew talking-head video with video-use
  • When captions render as boxes after video-use renders the final file
  • When English filler removal worked but Hebrew fillers like אֶה and כאילו remain
  • When editing a code-switched Hebrew/English tech tutorial with library names
  • When producing a Hebrew Reels or TikTok marketing video with video-use

Try These Prompts

Edit Hebrew launch video

Edit the files in this folder into a 5-minute Hebrew launch video with burned-in Hebrew captions. Use video-use with sub-force-style-hebrew from references/.

Fix caption boxes

My final video shows caption boxes. Run fc-list :lang=he, if missing install via scripts/install-hebrew-fonts.sh and re-render.

Remove Hebrew fillers

Read references/hebrew-filler-words.md and apply the list as a post-pass on my Scribe JSON. Add the timestamps to the cut-candidate set before building the EDL.

Reels captions

Split this video into 30-second clips in vertical 1080x1920 format using the vertical-social-he style from references/sub-force-style-hebrew.md.

Frequently Asked Questions

Changelog

v1.2.9

Docs: close 7 content gaps + fix doc/code contradiction. (1) captions-only-tuning.md added to Bundled Resources (was orphan). (2-4) Three new Troubleshooting entries: characters within words LTR not RTL (deliberate), no periods or question marks (default since 1.2.6), Scribe gap-recovery failure. (5) Fix major doc/code contradiction: Gotcha claiming "libass+SRT BiDi works correctly, do not pre-shape" was OPPOSITE of what the working script does. Rewrote as "double-reversal is DELIBERATE" with anti-regression guidance. (6) domain-checklist.md updated with 5 mitigated failure modes. (7) New references/quick-test.md: 10-second synthetic Hebrew clip recipe (say -v Carmit + ffmpeg) for ~$0.001 pipeline validation without burning real footage against free-tier quota.

May 16, 2026

v1.2.8

Docs: close 6 gaps surfaced during real-use validation. (1) Fix wrong ElevenLabs free-tier number (10K characters/month, not 10 hours, ≈ ~1 long Hebrew video/month). (2) Document the v1.2.7 auto gap-recovery feature. (3) Document the default punctuation-stripping behavior + how to disable. (4) --yes flag. (5) <video>.he.srt side-export. (6) New references/captions-only-tuning.md covering thresholds and scribe_v1/v2 split.

May 16, 2026

v1.2.7

Automatic Scribe gap + tail recovery: if Scribe drops 30s+ mid-file or 10s+ at the tail, captions-only.sh now detects the window, re-transcribes it in isolation with scribe_v2, and merges the recovered words back into the main transcript. Verified on a real 11:29 video where the last 53 seconds were silently untranscribed in a normal Scribe run.

May 16, 2026

Related Skills

skills-ilAuthor: skills-il
v1.2.0

Guide developers in integrating Israeli agritech tools and precision agriculture platforms including CropX (soil monitoring), Netafim GrowSphere (IoT irrigation), Taranis (crop intelligence), and the broader Israeli agritech ecosystem (approximately 600-750 companies per Start-Up Nation Central agrifoodtech). Use when user asks about agritech APIs, precision agriculture, smart irrigation, "hashkaya cham", crop monitoring, pest detection, Israeli agriculture tech, or needs to build farm management software. Covers irrigation optimization, pest detection, climate data integration, and Israeli agricultural context. Do NOT use for general gardening advice or non-agricultural IoT projects.

0.091,319
Claude CodeCursorGitHub Copilot+5
skills-ilAuthor: skills-il
v1.2.0PopularTrending

Benchmark and compare LLMs on Hebrew reasoning, comprehension, sentiment, translation, and Israeli cultural knowledge. Wraps the HuggingFace Open Hebrew LLM Leaderboard tasks (HeQ, HebrewSentiment, Hebrew Winograd, translation) plus DictaLM 3.0 benchmark tasks (Summarization, Nikud, Israeli Trivia) into a reproducible evaluation harness. Runs evals against Claude, GPT, Gemini, AI21 Jamba, DictaLM, Llama, and local HuggingFace models. Produces comparison scorecards in JSON and markdown. Use when choosing an LLM for a Hebrew product, answering procurement questions about Hebrew performance, validating a fine-tuned Hebrew model, or tracking Hebrew regressions after a model upgrade. Do NOT use for Arabic NLP, ASR benchmarking, or general English benchmarks.

0.0371,451
Claude CodeCursorCodex
skills-ilAuthor: skills-il
v1.2.0Popular

Guide Israeli startup operations including company formation, Innovation Authority grants, investment agreements, R&D tax benefits, and employee stock options (Option 102). Use when user asks about starting a company in Israel, IIA grants, "Innovation Authority", SAFE agreements (Israeli), convertible notes, Option 102, employee stock options in Israel, R&D tax benefits, preferred enterprise, Yozma 2.0, Delaware flip, or Israeli startup legal/financial setup. Do NOT use for non-Israeli company formation or international tax advice. Always recommend consulting with Israeli lawyer and accountant for binding decisions.

0.0572,020
Claude CodeCursorGitHub Copilot+5
Found an issue with this skill?

Use at your own risk. Terms of Use · Security

Want to build your own skill? Try the Skill Creator · Submit a Skill

Reviews (0)

No reviews yet. Be the first to write one!