메뉴
Animate a still image into a short video via OpenRouter. Default model google/veo-3.1-fast.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | openrouter-image2video |
| description | Animate a still image into a short video via OpenRouter. Default model google/veo-3.1-fast. |
Image + motion-prompt → video via OpenRouter. Default model: google/veo-3.1-fast.
Use this when you already have a strong still image and want to bring it to life with subtle motion (camera pan, parallax, gentle animation) while preserving the composition. For motion-from-scratch, use openrouter-text2video instead.
The script accepts either a remote image URL or a local image path; local files are base64-encoded and inlined into the request as a data URL.
Resolve TOOL_DIR = the directory containing this SKILL.md. Commands below use TOOL_DIR as a symbolic placeholder; replace it with the resolved, quoted path before running Bash.
export OPENROUTER_API_KEY=sk-or-v1-...
# From a local image you already generated with text2image
python3 TOOL_DIR/scripts/generate_video_from_image.py \
--image PROJECT_DIR/assets/teaser.png \
--prompt "slow parallax push-in, soft drift of ambient particles, no camera shake" \
--duration 5 \
--aspect-ratio 16:9 \
--download PROJECT_DIR/assets/teaser.mp4
# Or from a remote URL
python3 TOOL_DIR/scripts/generate_video_from_image.py \
--image-url "https://example.com/still.png" \
--prompt "subtle camera dolly forward, gentle depth-of-field shift" \
--download PROJECT_DIR/assets/scene.mp4
| Flag | Default | Description |
|---|---|---|
--prompt | required | Motion prompt — describe what should move and how |
--download | required | Output MP4 path |
--image | one of --image / --image-url required | Local image path (PNG/JPG); will be base64-encoded |
--image-url | one of --image / --image-url required | Remote image URL |
--model | google/veo-3.1-fast | Any OpenRouter image-to-video-capable model |
--duration | 5 | Seconds |
--aspect-ratio | 16:9 | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 |
--resolution | 720p | Model-dependent (e.g. 480p, 720p, 1080p) |
--frame-role | first | first or last — anchor frame role for the input image |
--generate-audio | off | Generate audio with video (if model supports) |
--poll-interval | 5 | Seconds between polls |
--max-wait | 600 | Max total wait time |
POST /api/v1/videos with body:
{
"model": "google/veo-3.1-fast",
"prompt": "...motion prompt...",
"aspect_ratio": "16:9",
"duration": 5,
"resolution": "720p",
"frame_images": [
{
"type": "image_url",
"frame_type": "first_frame",
"image_url": {"url": "data:image/png;base64,..." }
}
]
}
--frame-role first|last flag maps to frame_type: "first_frame"|"last_frame".GET /api/v1/videos/{id} every 5s until status == "completed"GET /api/v1/videos/{id}/content → raw MP4 bytesframe_images with roles first and last. The current script wires only one anchor frame; extend body["frame_images"] to add a second.Exhaustively profile a dataset and list ALL possible analyses — distributions, correlations, rankings, trends, group comparisons, anomalies. Reads detective.json for context. Runs after the Detective/Scout, before the Editor. Outputs analyst.json with ana_xx IDs and chart-ready data_tables.
Audit a generated Data2Story blog for build correctness across ALL modalities by ACTUALLY RENDERING it in a real headless browser (when available) — catching blank/0-width charts, broken/oversized media, desktop+mobile overflow, and dead interactions that source inspection cannot see — then having vision agents review the screenshots, playtesting every interactive, and walking the flagship capability contract. Plus VIEW every image for a wrong/AI-faked subject. Fixes layout in place; sends content-correctness problems back to the owning role. Use at Stage 6 after the Programmer assembles index.html and before the Critic; re-runs each revision round. Triggers: a built index.html exists, or you need to confirm charts/media/interactions actually render.
Design the MANDATORY cinematic scroll experience (every blog ships one): a full-bleed, scroll-driven background — primarily from the Scout's verified real imagery — plus the global motion choreography, so the narrative unfolds like a film as the reader scrolls. Owns the BACKGROUND layer + page-level motion (not the Designer's per-section visuals, not the Interaction centerpiece — it stages them inside the scroll). Outputs cinematographer.json (cin_xx scenes). There is no off path: a topic with no sourceable real imagery falls back to a generative-atmosphere or data-driven-spine background, never to a bare column.
Name the piece — re-write the masthead (headline + standfirst + kicker), every section title, and every figure/photo/table caption to a research-driven titling standard, killing the AI-tell patterns (the 'flat statement. flat counter-statement.' two-beat above all) a competent default falls into. Reads editor.md/json + analyst.json + the resolved topic_profile; writes copywriter.json — STRINGS ONLY (masthead{headline,standfirst,kicker}, items{edt_xx:{title}, des_xx:{caption}}), each backed by a real ana_*. Names, never edits: it touches no finding, no number, no data-* id, no layout — so the Verify layer is untouched and the Programmer renders the masthead + figcaptions from copywriter.json verbatim. Runs at Stage 3.5, after the Editor, before the Designer.
Review a finished Data2Story blog against the 5 quality rubric dimensions (visual_design, narrative_pacing, data_method_transparency, claim_data_alignment, insight_value), score each 1-7 with on-page evidence, and emit critic.json with pass/fail + targeted, surgical send-back instructions. Verifies every load-bearing claim/asset against its traceability chain before scoring; applies the caveat-survival, honest-accuracy, and third-party-attribution caps. Does NOT rewrite content — scores and sends back. Use at Stage 6.5 after the Auditor and after verify.py has produced verifier.json; re-runs each revision round. Triggers: a built index.html plus verifier.json exist, or you need to judge whether the article is actually good.
Generate text embeddings via OpenRouter using Qwen3-Embedding-8B.