One detail reads a person (cue → example → debate line)#

Template (timings assume duration 10; rescale for 5 or 15): 0–3s you can read someone by ONE cue; 3–8s one concrete example (user-chosen: e.g. how they treat service staff, how they handle a small promise, phone posture at dinner— no real non-public figures); 8–10s provocative summary that invites disagreement (e.g. THAT’S THE TELL—FIGHT ME toned playfully, not harassment). English subtitles with explicit timing. Default duration is 10 when unspecified.

Dependencies: WERYAI_API_KEY + Node.js 18+. scripts/video_gen.js + WERYAI_VIDEO_API.md. No other Cursor skills required.

Prerequisites#

• WERYAI_API_KEY must be set; Node 18+. Prefer public https image URLs; if the assembled scripts/video_gen.js supports local file paths, review/verify the script and explicitly consent before local read-and-upload to WeryAI.
• Non-empty model.

Security, secrets, and API hosts#

• WERYAI_API_KEY: Secret; requires.env / primaryEnv.
• API hosts fixed in video_gen.js; no URL env overrides.
• Local image handling disclosure: Prefer public https image URLs. If the assembled scripts/video_gen.js supports local file paths, it may read a local image and upload it to WeryAI to obtain a public URL; require review / verification and explicit consent before using that path.
• Higher assurance: Prefer a short-lived or isolated environment; review scripts/video_gen.js before production use. Verify whether the runtime can read local image files and upload them to WeryAI, and obtain explicit consent before using that path.

Prompt expansion (mandatory)#

• Duration: Default 10; 5 or 15 when allowed and requested.
• Timeline scaling: Rescale 0–3 / 3–8 / 8–10 proportionally; keep thesis → example → closer order.

Locked rhythm (example for duration 10):

• 0.0–3.0s — Thesis: e.g. ONE DETAIL EXPOSES THEIR LEVEL.
• 3.0–8.0s — Example beat (split into two on-screen lines if needed, e.g. 3–5.5 and 5.5–8.0) describing generic silhouettes, not identifiable people.
• 8.0–10.0s — Closer: strong opinion, playful fight-me energy— no slurs, no targeting protected groups.

Visuals: Close-ups hands, eyes, cafe table, door holds; shallow DOF; constant micro-motion.

Typography: Bold sans, outline, lower third; English on screen.

Negatives (Kling): illegible text, watermark.

Confirmation: Full prompt in table.

Niche checklist#

• Example must be filmable in B-roll, not a lecture.
• Avoid diagnosing real individuals or mental illness.

Workflow#

1. Text / image / multi-image ( SEEDANCE_2_0, ≤3).
2. Collect which cue the user cares about.
3. Expand → validate → confirm → wait.
4. [Video](url) for results.

CLI reference#

sh

node scripts/video_gen.js wait --json '{"model":"KLING_V3_0_PRO","prompt":"…","duration":10,"aspect_ratio":"9:16","generate_audio":true,"negative_prompt":"illegible text, watermark"}'
node scripts/video_gen.js wait --json '{"model":"SEEDANCE_2_0","prompt":"…","duration":10,"aspect_ratio":"9:16","resolution":"720p","generate_audio":true}'

Full reference: WERYAI_VIDEO_API.md.

Definition of done#

Video or error; timed English captions follow thesis / example / closer structure with ranges scaled to duration (or user-approved split).

Boundaries (out of scope)#

• No harassment, hate, or doxxing framing.
• Not professional psychology advice.
• Captions via prompt only; CLI details in resources/WERYAI_VIDEO_API.md.
• Multi-image only SEEDANCE_2_0, max 3.

Example prompts#

• Cue: how they treat waiters—cafe B-roll, playful closer
• Cue: punctuality on small promises—office hands clock
• Image: anonymous crowd silhouette URL → same beat structure
• Cue: how they apologize after a mistake—cafe reenactment, playful closer

---

Model and API constraints (frozen for this skill)#

Text-to-video#

model_keydurationsaspect_ratiosresolutionsaudionegative_promptprompt limitKLING_V3_0_PRO5, 10, 159:16, 1:1, 16:9_(omit)yesyes2000KLING_V3_0_STA5, 10, 159:16, 1:1, 16:9(omit)_yesyes2000SEEDANCE_2_05, 10, 159:16, 1:1, 16:9480p, 720pyesno2000

Image-to-video#

model_keydurationsaspect_ratiosresolutionsaudionegative_promptimage slotsKLING_V3_0_PRO5, 10, 159:16, 16:9, 1:1_(omit)yesyessingle imageKLING_V3_0_STA5, 10, 159:16, 16:9, 1:1(omit)_yesyessingle imageSEEDANCE_2_05, 10, 159:16, 1:1, 16:9480p, 720pyesnoup to 3 images

Recommended models#

Tiermodel_keyDefault duration⭐ BestKLING_V3_0_PRO10👍 GoodKLING_V3_0_STA10⚡ FastSEEDANCE_2_010 ( resolution: 720p)

Default parameters#

FieldValueaspect_ratio9:16duration10generate_audiotrueresolution720p (Seedance)negative_promptKling: illegible text, garbled captions, watermark

Text-to-video (primary)#

Collect cue → map to three windows → confirm → wait.

Image-to-video (optional)#

Reference still → emotional read with same structure.

Tips#

• Closer works when it’s obviously rhetorical, not cruel.
• Translate user Chinese brief to English captions unless they specify another on-screen language.

Packaged as one-detail-reads-person-video-gen.