Agent skill

Evidence-Grounded Video Writer

by liuluhaixiu in liuluhaixiu/DaVinci-AutoEdit-Agent

Writes video scripts, narration, titles and hooks from an approved media review, offers story directions first, and never invents footage, quotes or facts.

MITAuto-check passedMedia & Creative

Install Evidence-Grounded Video Writer

skills CLI
$ npx skills add liuluhaixiu/DaVinci-AutoEdit-Agent --skill viral-video-writer -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install liuluhaixiu/DaVinci-AutoEdit-Agent viral-video-writer --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/liuluhaixiu/DaVinci-AutoEdit-Agent.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/viral-video-writer .claude/skills/viral-video-writer && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
viral-video-writer
GitHub stars
484
Token cost
~711 tokens
SKILL.md length
326 words
Files
4 (incl. references)
Skills in repo
3
Repo updated
First seen
Licence
MIT

At a glance

Writes video scripts, narration, titles and hooks from an approved media review, offers story directions first, and never invents footage, quotes or facts.

  • Works in 8 steps: One memorable core idea → Two or three supporting ideas → Evidence available in the footage → …
  • Writing narration for a cut based on footage you have already reviewed
  • SKILL.md covers Intake, Develop Direction, Draft and Platform Adaptation, plus 1 more section
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

The agent works only from the approved brief and material review. It takes whatever topic, platform, audience, duration, language, tone and narration policy you already gave and asks for the rest. Before drafting it tests the idea against a short list that includes one core idea, evidence in the footage, audience tension, an emotional curve and the hook, turn, climax and aftertaste, then offers two or three distinct story directions and waits for your pick.

The draft uses short speakable sentences, leaves room for sound and silence, marks factual uncertainty, and ties each narrated beat to approved visual evidence or labels it as context you need to verify. It lists the scene sequence, narration intent, evidence references and rhythm notes, plus five title and hook options when publishing is in scope. The full draft appears in chat and is written to a file only after you approve. Platform versions are rewrites, not cuts, and two reference files hold a checklist and an ideation framework.

When your agent uses it

  • Writing narration for a cut based on footage you have already reviewed
  • Choosing between story angles for a short vertical video
  • Producing titles, hooks and captions for a specific platform and audience

Example prompts

  • “Using the approved review of the hiking footage, give me three story directions for a 60-second vertical video.”
  • “Write the narration for the documentary version in Spanish, and mark anything the footage does not show.”
  • “Draft five title and hook options for the product demo video aimed at developers.”

Requirements

  • An approved media review or factual brief to write from

Workflow steps

8 steps, taken from the first numbered list in SKILL.md.

  1. One memorable core idea
  2. Two or three supporting ideas
  3. Evidence available in the footage
  4. Audience tension or curiosity
  5. Emotional curve
  6. Point of view and voice
  7. Hook, turn, climax, and aftertaste
  8. Natural interaction prompt, when appropriate

What it can do on your machine

Read from SKILL.md and the folder at commit 93cf371. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Evidence-Grounded Video Writer loads about 711 tokens when it runs, and up to ~3.4k if it reads all its reference files. Until then it costs about 90 tokens; SKILL.md has 326 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~90
When it runs · the whole SKILL.md, loaded when a task matches
~711
With references · SKILL.md plus every file in references/, read only if the agent opens them
~3.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from liuluhaixiu/DaVinci-AutoEdit-Agent at commit 93cf371, republished under its MIT licence (© liuluhaixiu). 326 words, ~711 tokens.

Download SKILL.mdSave it as .claude/skills/viral-video-writer/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.
name
viral-video-writer
description
Evidence-grounded video script and narration writing for any topic, platform, audience, language, and duration. Use when Codex must turn an approved media review or factual brief into story directions, hooks, titles, narration, scene intent, emotional pacing, captions, or platform-specific video copy without inventing unavailable footage.

Viral Video Writer

Write from the approved brief and material review. Do not invent visuals, quotes, people, events, locations, or claims.

Intake

Use already supplied information. Ask only for missing topic, platform, audience, duration, language, tone, narration policy, and constraints.

Develop Direction

Before drafting, internally test:

  1. One memorable core idea
  2. Two or three supporting ideas
  3. Evidence available in the footage
  4. Audience tension or curiosity
  5. Emotional curve
  6. Point of view and voice
  7. Hook, turn, climax, and aftertaste
  8. Natural interaction prompt, when appropriate

Present 2-3 distinct story directions with tradeoffs. Obtain the user's choice before drafting the full script.

Draft

Match language density to the edit:

  • Do not narrate what the viewer can already understand.
  • Leave room for dialogue, action, natural sound, music, and silence.
  • Use short speakable sentences for voiceover.
  • Mark factual uncertainty instead of smoothing it over.
  • Tie each narrated beat to approved visual evidence or label it as a non-visual contextual statement requiring user verification.

Include:

  • selected core idea
  • scene/section sequence
  • narration or dialogue intent
  • visual evidence references
  • emotional and rhythm notes
  • five title/hook options when platform publishing is part of the request

Show the complete draft in chat. Write it only after explicit approval.

Platform Adaptation

Treat platform variants as rewrites, not truncations:

  • Long-form video: sustain questions, progression, callbacks, and breathing room.
  • Short-form vertical: immediate context, one idea, rapid evidence, clean payoff.
  • Documentary: protect ambiguity, observed detail, and factual confidence.
  • Tutorial/explainer: optimize comprehension, examples, and visual demonstrations.
  • Brand/event: preserve required facts without turning every line into a slogan.

Quality Bar

  • Specific beats generic.
  • Human cadence beats slogan density.
  • Emotional force must come from observed detail and structure.
  • A strong line should clarify the idea, not merely sound quotable.
  • The script must remain editable when TTS is disabled.

For the full writing checklist, read references/writing-checklist.md. For the original 11-dimension ideation framework and platform patterns, read references/viral-writing-framework.md; keep this skill's evidence and approval rules when that reference differs.

© liuluhaixiu, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 3 other files (references) in skills/viral-video-writer of liuluhaixiu/DaVinci-AutoEdit-Agent.

  • SKILL.md
  • agents/openai.yaml
  • references/viral-writing-framework.md
  • references/writing-checklist.md

Open the folder on GitHubat commit 93cf371

Compare with similar skills

Evidence-Grounded Video Writer next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Evidence-Grounded Video Writer compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Evidence-Grounded Video Writer this skillliuluhaixiu/DaVinci-AutoEdit-Agent484—~711Automated safety check: PassMIT
Seedance Interview1370998960-del/https-github.com-Emily2040-seedance-2.0241—~2.6kAutomated safety check: PassMIT
60-Second Cinematic Story Directorkaruvanan/MiniMax-H3-Director-Cut-Studio132—~1.8kAutomated safety check: PassCustom licence
Educational Content And How Tosocial-media-skills/skills134—~1.9kAutomated safety check: PassMIT
Video Script Writernicepkg/ai-workflow285—~2.2kAutomated safety check: PassMIT
Pitch Craft Timed Scriptsooiyeefei/ccc495—~1.7kAutomated safety check: PassMIT

Similar skills

  • Seedance Interview

    1370998960-del/https-github.com-Emily2040-seedance-2.0

    This skill should be used when the user has a vague Seedance 2.0 video idea and asks for creative guidance, story development, scene planning, a director interview, or help turning an undeveloped…

    241 GitHub stars~2.6k tokensUpdated 3 mo ago
    Media & CreativeAuto-check passed
  • 60-Second Cinematic Story Director

    karuvanan/MiniMax-H3-Director-Cut-Studio

    Adapts prose or rough story material into a roughly 60-second MiniMax H3 short with shot planning, Mandarin dialogue and voice-over, and a strong ending.

    132 GitHub stars~1.8k tokensUpdated today
    Media & CreativeAuto-check passed
  • Educational Content And How To

    social-media-skills/skills

    The teach-to-grow content type -- pick a genuinely useful topic, teach ONE thing completely, structure it as a clear lesson, keep it accurate, and route it to the right format.

    134 GitHub stars~1.9k tokensUpdated 9 days ago
    Media & CreativeAuto-check passed
  • Video Script Writer

    nicepkg/ai-workflow

    Write engaging video scripts for YouTube, TikTok, and other platforms.

    285 GitHub stars~2.2k tokensUpdated 8 mo ago
    Media & CreativeAuto-check passed
  • Writes or fixes spoken scripts that must fit a clock: demo voiceovers, deck narration and pitch speeches, with word budgets, storytelling rules and named trim points.

    495 GitHub stars~1.7k tokensUpdated 2 mo ago
    Writing & ContentAuto-check passed
  • Short Drama H3 Director

    karuvanan/MiniMax-H3-Director-Cut-Studio

    Develops vertical and short-form drama episodes for MiniMax H3 into causal stories, timed Shot Blocks, exact dialogue and Director Design JSON for an editable Timeline.

    132 GitHub stars~2.3k tokensUpdated today
    Media & CreativeAuto-check passed

More from liuluhaixiu/DaVinci-AutoEdit-Agent

  • DaVinci AutoEdit Agent

    liuluhaixiu/DaVinci-AutoEdit-Agent

    Guides an approval-gated video editing pipeline from raw footage to an audited DaVinci Resolve timeline, with scripting, optional TTS and a blueprint at each stage.

    484 GitHub stars~2.8k tokensUpdated 4 mo ago
    Auto-check: notes
  • DaVinci Resolve Editor

    liuluhaixiu/DaVinci-AutoEdit-Agent

    Edits DaVinci Resolve Studio projects and timelines safely through the davinci-resolve-mcp server or the Resolve Python API, with approval before writes and readback after.

    484 GitHub stars~852 tokensUpdated 4 mo ago
    Auto-check passed

Questions about Evidence-Grounded Video Writer

What does Evidence-Grounded Video Writer do?

Writes video scripts, narration, titles and hooks from an approved media review, offers story directions first, and never invents footage, quotes or facts. The agent works only from the approved brief and material review. It takes whatever topic, platform, audience, duration, language, tone and narration policy you already gave and asks for the rest.

When should I use Evidence-Grounded Video Writer?

Evidence-Grounded Video Writer fits situations like: writing narration for a cut based on footage you have already reviewed; choosing between story angles for a short vertical video; producing titles, hooks and captions for a specific platform and audience.

How do I install Evidence-Grounded Video Writer in Claude Code?

Run `npx skills add liuluhaixiu/DaVinci-AutoEdit-Agent --skill viral-video-writer -a claude-code`. Or copy the skill folder (skills/viral-video-writer in liuluhaixiu/DaVinci-AutoEdit-Agent) into .claude/skills/viral-video-writer in your project. Claude Code loads it when a task matches its description.

How do I install Evidence-Grounded Video Writer in Codex?

Run `npx skills add liuluhaixiu/DaVinci-AutoEdit-Agent --skill viral-video-writer -a codex`. Or copy the skill folder (skills/viral-video-writer in liuluhaixiu/DaVinci-AutoEdit-Agent) into .agents/skills/viral-video-writer in your project. Codex loads it when a task matches its description.

Can I use Evidence-Grounded Video Writer in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add liuluhaixiu/DaVinci-AutoEdit-Agent --skill viral-video-writer -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/viral-video-writer, .gemini/skills/viral-video-writer, .github/skills/viral-video-writer and .opencode/skills/viral-video-writer in your project.

What does Evidence-Grounded Video Writer need to run?

SKILL.md names no scripts, command-line tools or credentials: Evidence-Grounded Video Writer is instructions for the agent only. Our summary lists: An approved media review or factual brief to write from.

Does Evidence-Grounded Video Writer access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Evidence-Grounded Video Writer safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Evidence-Grounded Video Writer use?

Evidence-Grounded Video Writer is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Evidence-Grounded Video Writer use?

About 711 tokens (SKILL.md is roughly 2.8k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.7k tokens, read only when the agent opens those files.

What are the alternatives to Evidence-Grounded Video Writer?

Skills that share tags, products or a category with Evidence-Grounded Video Writer: Seedance Interview (1370998960-del/https-github.com-Emily2040-seedance-2.0, 241 stars), 60-Second Cinematic Story Director (karuvanan/MiniMax-H3-Director-Cut-Studio, 132 stars), Educational Content And How To (social-media-skills/skills, 134 stars) and Video Script Writer (nicepkg/ai-workflow, 285 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Evidence-Grounded Video Writer?

liuluhaixiu (a GitHub user) maintains it in liuluhaixiu/DaVinci-AutoEdit-Agent, which has 484 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on June 12, 2026.

Source: liuluhaixiu/DaVinci-AutoEdit-Agent on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.