Agent skill

Moonvalley Marey

by calesthio in calesthio/generative-media-skills

Produce video with Moonvalley's Marey model family (Marey Realism v1.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data.

MITAuto-check passedMedia & Creative

Install Moonvalley Marey

skills CLI
$ npx skills add calesthio/generative-media-skills --skill moonvalley-marey -a claude-code

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

GitHub CLI
$ gh skill install calesthio/generative-media-skills moonvalley-marey --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/calesthio/generative-media-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/providers/video-generation/moonvalley-marey .claude/skills/moonvalley-marey && 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
moonvalley-marey
GitHub stars
197
Token cost
~5.2k tokens
SKILL.md length
2,627 words
Files
2
Skills in repo
26
Repo updated
First seen
Licence
MIT

At a glance

Produce video with Moonvalley's Marey model family (Marey Realism v1.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data.

  • Works in 5 steps: Camera first. Name a concrete move and… → Scale / perspective. Anchor size with… → Core visual / action. The single primary… → …
  • A task asks for Marey
  • SKILL.md covers Company and model status…, Documented capabilities and…, The licensed-data value… and Access routes (verified…, plus 9 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Moonvalley Marey is an agent skill from calesthio/generative-media-skills. Produce video with Moonvalley's Marey model family (Marey Realism v1.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data. Use this skill when a task asks for Marey or Moonvalley specifically, when a brief demands brand-safe / legally reviewed AI video for commercial or studio work, or when the request needs Marey's director-style controls (camera trajectory, motion transfer, pose transfer, keyframing, reference conditioning). It covers what "commercially…

Its SKILL.md is about 5.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 1 other file (for example `EVAL.md`).

It sits in Media & Creative, covering AI video generation, Policy and terms drafting and Diffusion and image models. It works with fal and ComfyUI. The repository describes itself as: Research-backed agent skills and tools for premium image, video, audio, voice, and generative media production across AI coding assistants. The licence is MIT.

When your agent uses it

  • A task asks for Marey
  • Moonvalley specifically
  • A brief demands brand-safe / legally reviewed AI video for commercial
  • The request needs Mareys director-style controls (camera trajectory

Example prompts

  • “commercially safe”
  • “best video model”
  • “/moonvalley-marey”

Workflow steps

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

  1. Camera first. Name a concrete move and angle — dolly-in, aerial sweep,
  2. Scale / perspective. Anchor size with real-world references (monolith,
  3. Core visual / action. The single primary subject and action. Keep it to
  4. Environment, in layers. Separate foreground / middle ground / background;
  5. Lighting and technical spec last. Lens (35mm), sensor, film stock, noir

What it can do on your machine

Read from SKILL.md and the folder at commit 8c85352. 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

Moonvalley Marey loads about 5.2k tokens when it runs. Until then it costs about 230 tokens; SKILL.md has 2,627 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~230
When it runs · the whole SKILL.md, loaded when a task matches
~5.2k

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 calesthio/generative-media-skills at commit 8c85352, republished under its MIT licence (© calesthio). 2,627 words, ~5,167 tokens.

Download SKILL.mdSave it as .claude/skills/moonvalley-marey/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
moonvalley-marey
description
Produce video with Moonvalley's Marey model family (Marey Realism v1.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data. Use this skill when a task asks for Marey or Moonvalley specifically, when a brief demands brand-safe / legally reviewed AI video for commercial or studio work, or when the request needs Marey's director-style controls (camera trajectory, motion transfer, pose transfer, keyframing, reference conditioning). It covers what "commercially safe" really means versus the actual terms of service, access routes (Moonvalley web / Voyager, the fal API, ComfyUI), pricing, capability limits, prompt and reference strategy, where Marey fits production versus where other models win, iteration, and quality review. Do not use it for audio generation, for photoreal talking-head dialogue, or as a generic "best video model" default.

Moonvalley Marey

Marey is Moonvalley's generative video model, aimed at professional filmmakers, studios, and brand teams that need AI video they can put through legal review. Its differentiator is not raw fidelity — it is provenance: Moonvalley markets Marey as trained exclusively on owned or licensed footage, and pairs it with a set of director-style controls. This skill helps an agent decide when Marey is the right tool, use it correctly, and — critically — represent its legal posture accurately rather than repeating marketing.

Verification date for all volatile facts below: 2026-07-10. Re-check version, pricing, endpoints, and the terms of service before relying on any dated claim.

Company and model status (verified 2026-07-10)

  • Company: Moonvalley (moonvalley.com), an "imagination research company." Marey was built with the animation studio Asteria. Moonvalley raised an additional $84M in July 2025 (investors incl. General Catalyst, CAA, Comcast Ventures, CoreWeave, Khosla Ventures, Y Combinator), reportedly ~$154M total. The company is active and funded — not shut down or pivoted. [businesswire, variety, techcrunch — see Sources]
  • Model: The publicly available production model is Marey Realism v1.5. It first entered public availability July 2025 after an earlier limited release (March 2025). [techcrunch, fal]
  • Platform: Moonvalley also markets Voyager, a higher-end platform/UI that exposes Marey's fuller control set for studio workflows. [businesswire]
  • Named after Étienne-Jules Marey, a chronophotography pioneer — a signal of the "cinematography-first" brand positioning.

Treat "v1.5" as volatile. Moonvalley has publicly promised further controls (lighting, deeper object trajectories, character libraries); a newer point release may exist by the time you read this.

Documented capabilities and hard limits

These are documented on the fal model pages and Moonvalley/partner docs (verified 2026-07-10). Distinguish them from marketing claims below.

PropertyValueNote
Output resolutionNative 1080p (default 1920×1080)Trained on native 1080p; other dimensions selectable
Frame rate24 fpsCinema-standard cadence
Clip duration5s or 10s per generation (v1.5 on fal); shot-extension to lengthenOriginal 2025 launch capped at ~5s; v1.5 adds 10s + extension
AudioNoneMarey is silent video only — no native dialogue, SFX, or music. Add audio in post
Training content"Owned or fully licensed," native 1080p, no user-generated content (first-party claim)See legal section

Consequences an agent must plan around:

  • No audio track. Any brief needing lip-synced dialogue or sound must layer audio separately (ElevenLabs, a DAW, licensed music). If the brief's value is in synced sound, Marey is not the model — Veo 3 generates audio natively.
  • Short clips. Narrative beats must be built from 5–10s shots and assembled in an NLE. Plan a shot list, not a single long generation.
  • 1080p ceiling. For 4K delivery you upscale in post; Marey does not natively deliver 4K.

The licensed-data value proposition — and what the terms actually say

This is the single most important thing to get right, because it is where an agent most easily misleads a user.

The first-party claim (label it as a claim)

Moonvalley markets Marey as "the world's first commercially safe video model ... trained only on licensed, high-resolution footage. No scraped content. No user submissions." Reporting indicates roughly 80% of training footage came from filmmakers/agencies who intentionally licensed B-roll, with archive material secured through partnerships (e.g. Vimeo); the remainder from other owned/licensed sources. This is a first-party claim about training-data provenance, corroborated by press but not independently audited. Present it as "Moonvalley states / claims," never as an established fact. [moonvalley, cined, techcrunch]

Why provenance matters

The plausible legal benefit is at the input side: a model trained only on cleared footage is far less exposed to the training-data copyright suits now facing scraped-data competitors. For a risk-averse studio or brand, that reduces one specific category of risk — the "was this model built on my/others' stolen work" question — and that is a real, defensible reason to choose Marey.

What "commercially safe" does NOT mean — read the ToS

Do not equate "trained on licensed data" with "you are indemnified." The actual consumer Terms of Service (help.moonvalley.com, verified 2026-07-10) run the other way:

  • The user indemnifies Moonvalley, not the reverse. Users agree to "defend, indemnify, and hold harmless" Moonvalley from claims arising out of the user's use of the service, their content, or infringement of others' rights — including paying defense costs. [moonvalley ToS]
  • Liability is capped at the greater of $500 or the amount the user paid. Moonvalley disclaims indirect/consequential/special damages. [moonvalley ToS]
  • Reporting (TechCrunch, citing a Moonvalley spokesperson) indicates Moonvalley provides indemnification for its data-collection contractors — not a blanket output indemnity to end users. This is materially different from Adobe Firefly, which offers enterprise customers indemnification for using its video model. [techcrunch]
  • Moonvalley has said it intends to offer a user indemnity policy and content-removal/data-deletion options, and it markets bespoke enterprise / studio agreements (including custom fine-tunes on partners' IP). Any real output indemnity for a specific customer would live in a negotiated enterprise contract, not the public ToS. [techcrunch, deadline]

Agent rule: When a user cites "commercially safe" as a reason to skip legal review, correct the framing. The honest position is: Marey lowers training-data provenance risk and is a strong candidate for brand-safe work, but the standard terms shift indemnification onto the user and cap liability — so studios must still run legal review, and any indemnity guarantee must be secured in an enterprise agreement, in writing. Never tell a user their Marey output is "legally cleared" or "guaranteed" on the strength of marketing copy alone.

Access routes (verified 2026-07-10)

  • Moonvalley web app / Voyager (moonvalley.com): the first-party product; exposes the fullest control set and the credit store.
  • fal API (fal.ai/models/moonvalley/marey/...): programmatic access to t2v, i2v, motion-transfer, and pose-transfer endpoints. Best route for agent automation and pipelines.
  • ComfyUI: Marey Realism v1.5 is referenced for node-based workflows.
  • Third-party creative platforms: Marey has appeared via partners (e.g. Adobe integrations, Scenario). Availability and feature parity vary; verify per platform.
fal API parameters (v1.5 t2v/i2v, verified 2026-07-10)
  • prompt (required): descriptive text; minimum ~50 words recommended.
  • image_url (i2v only, required): the starting frame.
  • negative_prompt (optional): ships with sensible defaults; customizable.
  • seed (optional): default 9; use -1 for random. Fix the seed to iterate one variable at a time; randomize to explore.
  • dimensions (optional): resolution/aspect via dropdown; default 1920×1080.
  • duration (optional): 5s or 10s.

Pricing (verified 2026-07-10)

  • fal, per generation: $1.50 for 5s, $3.00 for 10s. [fal]
  • Moonvalley web credit packs: $14.99 / 100 credits, $34.99 / 250 credits, $149.99 / 1,000 credits. A 5s request is billed like the $1.50 tier, 10s like $3.00. [techcrunch]
  • Custom fine-tuning is available to studios (enterprise pricing, not public).

Budget implication: iteration is the real cost. At ~$1.50–$3.00 per attempt and Marey's inconsistency (below), a single usable 5s shot may take several generations. Estimate 3–8 attempts per keeper shot when quoting a budget.

Filmmaker controls (feature names verbatim, verified 2026-07-10)

Marey's pitch is directorial control, not one-shot prompting. Feature names as Moonvalley lists them:

  • Camera Control — "Create cinematic camera moves using just a single image." Near-360° moves, handheld/dolly simulation.
  • Motion Transfer — "Pull motion from any reference video and apply it to new subjects." Use a reference clip as a motion template; restyle the visuals.
  • Pose Control / Pose Transfer — drive human movement, gesture, and body motion from reference images, videos, or performance inputs.
  • Trajectory Control — "Draw the path. Define the action." Direct object/ motion paths.
  • Keyframing — "Upload multiple reference images and place them on a timeline."
  • Reference — "Use separate reference images for each character or object" (per-subject conditioning / character consistency).
  • Shot Extension — "Seamlessly extend video duration" beyond a single generation.
  • Multi-layer spatial composition — independent control of foreground, middle ground, background (Marey is marketed as "3D-aware").

Heuristic: reach for these controls precisely because Marey's blind text-to-video is inconsistent. Motion transfer and pose transfer replace fragile prompted motion with a real reference, which is where Marey earns its place — previz, restyling a plate, or matching a specific camera move.

Prompt strategy

Marey rewards long, structured, cinematographer's prompts far more than terse ones. Moonvalley's own guidance recommends at least ~50 words and a layered structure. A documented formula (Moonvalley help center):

[Camera movement] + [Scale / perspective] + [Core visual] + [Environmental details] + [Lighting / technical specs]

Practical construction:

  1. Camera first. Name a concrete move and angle — dolly-in, aerial sweep, low-angle push, over-the-shoulder, plunge, orbit.
  2. Scale / perspective. Anchor size with real-world references (monolith, canyon, building) so the model stages depth.
  3. Core visual / action. The single primary subject and action. Keep it to one clear action — Marey degrades when asked to juggle multiple simultaneous actions or props.
  4. Environment, in layers. Separate foreground / middle ground / background; this plays to the multi-layer composition strength.
  5. Lighting and technical spec last. Lens (35mm), sensor, film stock, noir lighting, motion blur, shallow focus, golden hour.

Negative prompt: keep the shipped defaults and add specific failure terms you observe (e.g. warped hands, flicker, morphing, extra fingers, plastic skin).

Don't: pack a paragraph with several characters each doing different things; ask for readable text/logos; expect synced lips or spoken words.

Show full SKILL.md (1,163 more words)Show less

Where Marey fits production — and where other models win

Choose Marey when:

  • The client is a brand or studio with legal review and training-data provenance is a gating requirement.
  • The shot is a single-subject, controlled piece — product hero shot, slow aerial/establishing, an environment where motion is minimal, or a stylized/2D look that hides temporal artifacts.
  • You have a reference to drive motion or pose (motion/pose transfer), or a previz need where directorial control matters more than final fidelity.
  • You are combining Marey with its own ecosystem (Voyager controls, keyframing, shot extension) rather than expecting one clean text-to-video generation.

Prefer another model when:

  • The brief needs native audio / dialogue → Veo 3.
  • The brief needs sustained, complex, multi-character motion, crowd choreography, or reliable photoreal talking heads → Marey's temporal consistency is weak; competitors (Veo, Kling, Seedance, Sora-class) handle these more reliably per independent hands-on reviews.
  • You need maximum prompt adherence on the first try at low cost — Marey's adherence is documented as inconsistent, so iteration cost is high.
  • Provenance is not a requirement and you simply want the best-looking result fast.

This is a first-party-claim-vs-independent-review split: Marey's licensing story is genuinely differentiated, but a hands-on practitioner review (Curious Refuge, 2025) scored its output quality low (~3/10) versus peers, citing temporal instability, collapsing hands/gestures, flicker on camera moves, and plastic skin. Treat that as one disclosed-method practitioner report, not a benchmark — but it aligns with the "brittle under motion" pattern. The honest summary: Marey buys legal peace of mind and directorial control, often at a cost in raw motion fidelity.

Iteration workflow

  1. Lock the concept as a shot list, each shot ≤10s. Marey builds sequences from short clips, not long takes.
  2. Storyboard / previz cheaply with text-to-video, then commit budget.
  3. Fix the seed and change one variable per generation (prompt clause, duration, dimension). Randomize only when exploring fresh directions.
  4. Replace fragile prompted motion with references. If a camera move or gesture won't hold, drive it with motion transfer / pose transfer / camera control instead of more adjectives.
  5. Prefer i2v for control. Generate or source a strong first frame (an image model or a real still), then animate it — this pins composition and identity far better than blind t2v.
  6. Use keyframing and shot extension to stitch beats and lengthen usable shots rather than regenerating whole sequences.
  7. Finish in post: upscale toward delivery resolution, color, and add all audio. Marey outputs a silent 1080p/24fps plate.

Quality review criteria

Before accepting a Marey shot, check specifically for its known failure modes:

  • Temporal consistency: do props/objects persist across the whole clip, or vanish/morph? Watch the full duration, not the first frame.
  • Hands and gestures: fingers and hand articulation are a frequent collapse point — scrutinize any gesture.
  • Camera-move flicker: pans, orbits, and dollies expose jitter/flicker; inspect motion segments frame-by-frame.
  • Skin and faces: check for plastic/waxy skin and mannequin-stiff performance in any human shot.
  • Prompt adherence: did every requested element appear and survive? Multi- element prompts commonly drop details.
  • Delivery fit: 1080p and 24fps — confirm this matches the deliverable spec before committing (e.g. a 30fps social spec or 4K broadcast spec needs a plan).
  • Legal: confirmed brand/likeness/trademark clearance for anything you fed in (references, first frames) — the licensed-training claim does not cover your inputs, and the ToS makes those your responsibility.

Failure modes and repairs

SymptomLikely causeRepair
Props/details vanish mid-clipOverloaded prompt, too many simultaneous elementsSimplify to one core action; use i2v to pin the scene; add missing item to negative-of-absence via reference
Warped hands / collapsing gesturesKnown weak spot in fine articulationAvoid tight hand close-ups; use pose transfer with a clean reference; hide hands or keep them still
Flicker on camera movementMotion instabilityUse Camera Control from a single image instead of prompted motion; slow the move; shorten the clip
Plastic/stiff human performancePhotoreal-dialogue is not a strengthReconsider model for talking heads; use stylized look; drive performance via pose/motion transfer
Weak prompt adherenceTerse or unstructured promptRewrite to ≥50 words in the camera→scale→action→environment→lighting structure
Needs soundMarey is silentAdd audio in post; if synced audio is the point, switch to an audio-native model

Examples (illustrative — adapt, do not copy verbatim)

Example A — brand-safe product hero (Marey's sweet spot)
  • Intent: 5s hero shot of a perfume bottle for a brand ad going through legal review; provenance matters, motion is minimal.
  • Route: fal i2v. Start from a rendered product still to lock the label.
  • Prompt (~55 words): "Slow dolly-in, eye-level. A single frosted-glass perfume bottle stands like a monolith on wet black stone, foreground droplets catching light, soft gradient studio backdrop behind. Warm key from camera-left, cool rim light. Shallow depth of field, 50mm lens, gentle motion blur, high-contrast luxury lighting, pristine reflections, cinematic product photography."
  • Params: duration: 5, seed: 9 (fixed to iterate), default dimensions.
  • Why: single subject, near-static, controlled light — plays to strengths; i2v pins the label so it doesn't morph.
  • Likely failures: minor reflection flicker; label text warping if relied on the model instead of the input still.
  • Review: confirm the label stays legible and stable across all 5s; confirm the brand cleared the still they supplied — the training claim doesn't cover their asset.
Example B — motion transfer for previz (control over fidelity)
  • Intent: restyle a rehearsal handheld clip into a noir look for a director's previz.
  • Route: fal motion-transfer — reference video supplies the camera/subject motion; prompt supplies style.
  • Prompt: describe the target look and lighting only ("noir, hard chiaroscuro key, rain-slick street, 35mm, heavy shadow, desaturated") and let the reference carry motion.
  • Why: replaces fragile prompted camera motion with a real move — the reliable way to get a specific camera trajectory out of Marey.
  • Failures: gesture detail may still collapse; keep it a previz, not final.
Example C — when to say no
  • Request: "30s photoreal two-person dialogue scene with synced speech."
  • Correct response: Marey is the wrong tool — no native audio, 5–10s clips, and weak photoreal-dialogue consistency. Recommend an audio-native model (Veo 3) for the talking coverage, or Marey only for cutaway/establishing plates with audio and dialogue built separately. Do not promise a single Marey generation can deliver this.

Sources (verified 2026-07-10)

First-party (Moonvalley / partners):

  • Moonvalley product & Marey page — moonvalley.com, /marey (capabilities, control names, "commercially safe" claim)
  • Moonvalley Terms of Service — help.moonvalley.com/en/articles/10207682 (user indemnification, ~$500 liability cap)
  • Moonvalley prompting guidance — help.moonvalley.com (≥50-word prompt formula)
  • fal model pages — fal.ai/models/moonvalley/marey/t2v and /i2v (v1.5, pricing $1.50/5s & $3.00/10s, 1920×1080 default, seed default 9, endpoints)
  • fal blog — blog.fal.ai (v1.5 endpoints incl. motion/pose transfer)
  • Businesswire — $84M raise, Voyager, "first fully-licensed" claim

Secondary / reporting (labeled):

  • TechCrunch (2025-03-12, 2025-07-08, 2025-04-07) — public availability, indemnity-for-contractors-not-users distinction vs Adobe Firefly, pricing tiers
  • CineD, VentureBeat, Deadline, PetaPixel, Variety, SiliconANGLE — commercial- safety framing, ~80% licensed figure, Vimeo partnership, funding

Independent practitioner report (disclosed method, single reviewer):

  • Curious Refuge, "Marey by Moonvalley — An Honest Review" (2025) — hands-on quality scoring (~3/10), temporal/motion/adherence weaknesses vs Veo 3 / Seedance. Treat as one practitioner report, not a benchmark.

Re-verify every dated fact — versions, endpoints, pricing, and especially the terms of service — before relying on it.

© calesthio, 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 1 other file in skills/providers/video-generation/moonvalley-marey of calesthio/generative-media-skills.

  • SKILL.md
  • EVAL.md

Open the folder on GitHubat commit 8c85352

Compare with similar skills

Moonvalley Marey 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.

Moonvalley Marey compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Moonvalley Marey this skillcalesthio/generative-media-skills197—~5.2kAutomated safety check: PassMIT
Seedance 2 5calesthio/OpenMontage66k—~3.2kAutomated safety check: PassAGPL-3.0
RunninghubHM-RunningHub/OpenClaw_RH_Skills142—~1.6kAutomated safety check: PassApache-2.0
H3 Videoagent-next/video-agent120—~2.9kAutomated safety check: PassApache-2.0
Open Videoagent-next/video-agent120—~3.2kAutomated safety check: PassApache-2.0
ComfyUI Local DriverSlavaSexton/ComfyUI-Agent-Kit105—~12kAutomated safety check: PassApache-2.0

Similar skills

  • Seedance 2 5

    calesthio/OpenMontage

    Generate 4-30 second cinematic video with ByteDance Seedance 2.5 through fal.ai, Volcengine Ark, Runway, or ComfyUI Partner Nodes.

    66k GitHub stars~3.2k tokensUpdated 7 days ago
    Media & CreativeAuto-check passed
  • Runninghub

    HM-RunningHub/OpenClaw_RH_Skills

    Generate images, videos, audio, and 3D models via RunningHub API (420 endpoints) and run any RunningHub AI Application (custom ComfyUI workflow) by webappId.

    142 GitHub stars~1.6k tokensUpdated 1 mo ago
    Media & CreativeAuto-check passed
  • H3 Video

    agent-next/video-agent

    OpenVideo skill (v0.1.0): generate high-quality local video with the OpenVideo product (MiniMax H3 backend).

    120 GitHub stars~2.9k tokensUpdated yesterday
    Media & CreativeAuto-check passed
  • Open Video

    agent-next/video-agent

    Generate, edit, or direct videos via open-source models (MiniMax H3 baseline; Wan2.2 / LTX future).

    120 GitHub stars~3.2k tokensUpdated yesterday
    Media & CreativeAuto-check passed
  • ComfyUI Local Driver

    SlavaSexton/ComfyUI-Agent-Kit

    Drives a local ComfyUI install over its HTTP API to generate and edit images, video and audio, with per-model prompt recipes and workflow guidance.

    105 GitHub stars~12k tokensUpdated 1 mo ago
    Media & CreativeAuto-check passed
  • Minimax H3

    calesthio/OpenMontage

    Generate MiniMax H3 (Hailuo 3.0) video through the official MiniMax v2 API, fal.ai, Runway, ComfyUI Partner Nodes, or local open weights in ComfyUI.

    66k GitHub stars~580 tokensUpdated 7 days ago
    Media & CreativeAuto-check passed

More from calesthio/generative-media-skills

All 26 skills in this repo
  • 3D Asset Production

    calesthio/generative-media-skills

    A skill your agent uses to turn generated, captured, scanned, or modeled 3D output into production-ready standalone assets for DCC, real-time engine, web, or interchange delivery.

    197 GitHub stars~9.7k tokensUpdated 2 mo ago
    Auto-check passed
  • Audio Mixing Mastering

    calesthio/generative-media-skills

    Provider-independent audio mixing and mastering direction for AI agents finishing generated videos, ads, trailers, explainers, podcasts, recuts, avatar clips, music videos, documentaries, and social…

    197 GitHub stars~7k tokensUpdated 2 mo ago
    Auto-check passed
  • Captions Media Accessibility

    calesthio/generative-media-skills

    Provider-independent captions and media accessibility direction for AI agents producing or finishing generated videos, ads, social clips, explainers, avatar videos, documentaries, podcasts/video…

    197 GitHub stars~6.9k tokensUpdated 2 mo ago
    Auto-check passed
  • Comfyui Media Workflows

    calesthio/generative-media-skills

    Provider-independent production workflow for agents assembling, auditing, executing, and handing off ComfyUI node-graph workflows for image, video, upscale, inpaint, conditioning, and batch media…

    197 GitHub stars~8.5k tokensUpdated 2 mo ago
    Auto-check passed
  • Ffmpeg Media Finishing

    calesthio/generative-media-skills

    Provider-independent FFmpeg finishing workflow for AI agents preparing generated or edited media deliverables.

    197 GitHub stars~8.2k tokensUpdated 2 mo ago
    Auto-check passed
  • Generated Media QA

    calesthio/generative-media-skills

    Provider-independent quality assurance for AI-generated and AI-assisted media.

    197 GitHub stars~8k tokensUpdated 2 mo ago
    Auto-check passed

Works with

Questions about Moonvalley Marey

What does Moonvalley Marey do?

Produce video with Moonvalley's Marey model family (Marey Realism v1.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data. Moonvalley Marey is an agent skill from calesthio/generative-media-skills.5) — a filmmaker-oriented, 1080p/24fps generative video model marketed as trained exclusively on licensed data.

When should I use Moonvalley Marey?

Moonvalley Marey fits situations like: A task asks for Marey; moonvalley specifically; A brief demands brand-safe / legally reviewed AI video for commercial; the request needs Mareys director-style controls (camera trajectory.

How do I install Moonvalley Marey in Claude Code?

Run `npx skills add calesthio/generative-media-skills --skill moonvalley-marey -a claude-code`. Or copy the skill folder (skills/providers/video-generation/moonvalley-marey in calesthio/generative-media-skills) into .claude/skills/moonvalley-marey in your project. Claude Code loads it when a task matches its description.

How do I install Moonvalley Marey in Codex?

Run `npx skills add calesthio/generative-media-skills --skill moonvalley-marey -a codex`. Or copy the skill folder (skills/providers/video-generation/moonvalley-marey in calesthio/generative-media-skills) into .agents/skills/moonvalley-marey in your project. Codex loads it when a task matches its description.

Can I use Moonvalley Marey 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 calesthio/generative-media-skills --skill moonvalley-marey -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/moonvalley-marey, .gemini/skills/moonvalley-marey, .github/skills/moonvalley-marey and .opencode/skills/moonvalley-marey in your project.

What does Moonvalley Marey need to run?

SKILL.md names no scripts, command-line tools or credentials: Moonvalley Marey is instructions for the agent only.

Does Moonvalley Marey 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 Moonvalley Marey 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 Moonvalley Marey use?

Moonvalley Marey 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 Moonvalley Marey use?

About 5.2k tokens (SKILL.md is roughly 21k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Moonvalley Marey?

Skills that share tags, products or a category with Moonvalley Marey: Seedance 2 5 (calesthio/OpenMontage, 66k stars), Runninghub (HM-RunningHub/OpenClaw_RH_Skills, 142 stars), H3 Video (agent-next/video-agent, 120 stars) and Open Video (agent-next/video-agent, 120 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Moonvalley Marey?

calesthio (a GitHub user) maintains it in calesthio/generative-media-skills, which has 197 GitHub stars. The repository holds 26 skills in this directory. The repository was last updated on July 14, 2026.

Source: calesthio/generative-media-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.