Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/belagum/promptcut/capcut-draftnpx skills add Belagum/promptcut --skill capcut-draftgit clone --depth 1 https://github.com/Belagum/promptcutWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/belagum/promptcut/capcut-draft)<a href="https://agentmods.dev/skills/belagum/promptcut/capcut-draft"><img src="https://agentmods.dev/badge/skills/belagum/promptcut/capcut-draft.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00071 | $0.02099 |
| Opus 5 | $0.00036 | $0.01050 |
| Sonnet 5 | $0.00014 | $0.00420 |
| Haiku 4.5 | $0.00007 | $0.00210 |
Grade A, and why
capcut-draft scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CapCut drafts
PromptCut writes a real CapCut project folder. The person opens CapCut, sees the timeline with every clip, effect and subtitle in place, tweaks it, and exports from CapCut. Nothing is rendered here.
Requirements: pip install pycapcut, CapCut desktop, and the drafts folder.
Run capcut-drafts to find it; if it cannot be found, ask the person for
CapCut → Settings → Draft location, then
config --set capcut_drafts_dir="<path>".
CapCut must be restarted to notice a new draft. Do not write to a draft that is open. Only unencrypted drafts can be read back; newer CapCut versions encrypt them, which affects loading templates, not creating drafts.
Two ways in
From an existing storyboard:
python "${CLAUDE_PLUGIN_ROOT}/scripts/promptcut.py" capcut-from-plan --plan plan.json
Stills with their Ken Burns as scale/position keyframes, overlay titles as PNG
cards on a second video track, voiceover (a sentence group as one whole take),
music with the ducking curve as volume keyframes, sfx, and SRT subtitles - one
track each. It shares the timeline model with the Premiere/Resolve/VEGAS export
(nle-export skill), so all four editors see the same cut. --use-clips puts the
already rendered motion clips on the timeline instead of stills.
--transitions adds CapCut transitions - they eat time from both neighbours, so
the video drifts against the voiceover track; leave it off when sync matters.
Anything else - write a spec and build it:
python "${CLAUDE_PLUGIN_ROOT}/scripts/promptcut.py" capcut-build --spec spec.json
Add --dump out/draft_content.json to inspect the result without touching the
drafts folder.
Spec schema
{
"name": "my_draft",
"size": [1080, 1920],
"fps": 30,
"replace": true,
"tracks": [
{"type": "video", "name": "main", "index": 0, "mute": false, "segments": [
{"file": "shot1.png",
"start": 0, "duration": 4,
"source_in": null, "source_duration": null,
"speed": 1.0, "volume": 1.0,
"clip": {"scale": 1.05, "scale_x": null, "scale_y": null,
"x": 0, "y": 0, "rotation": 0, "alpha": 1,
"flip_h": false, "flip_v": false},
"animation_in": "zoom_in", "animation_in_dur": 0.8,
"animation_out": "fade_out", "animation_group": null,
"transition": "dissolve", "transition_dur": 0.5,
"filter": "Peach_Fuzz", "filter_intensity": 60,
"mask": {"type": "circle", "cx": 0, "cy": 0, "size": 0.8,
"rotation": 0, "feather": 0.2, "invert": false,
"rect_width": null, "round_corner": null},
"effects": ["Beat_Shots", "char:Pop_Out"],
"background": {"type": "blur", "blur": 0.0625},
"keyframes": [{"prop": "scale_x", "at": 0, "value": 1.0},
{"prop": "scale_x", "at": 4, "value": 1.2}]}]},
{"type": "audio", "name": "vo", "segments": [
{"file": "vo.mp3", "start": 0.15, "duration": 3.2, "volume": 1.0,
"speed": 1.0, "fade_in": 0.2, "fade_out": 0.4, "effect": null,
"keyframes": [{"at": 0, "value": 1.0}, {"at": 3, "value": 0.4}]}]},
{"type": "text", "name": "titles", "segments": [
{"text": "BIG TITLE", "start": 0.2, "duration": 2.5,
"font": "Anton", "size": 9, "color": "#FFFFFF", "bold": true,
"align": 1, "alpha": 1, "letter_spacing": 0, "line_spacing": 0,
"auto_wrap": true, "max_line_width": 0.82,
"border": {"color": [0, 0, 0], "width": 45, "alpha": 1},
"background": {"color": "#000000", "alpha": 0.5, "round": 0.2},
"clip": {"y": -0.3},
"anim_in": "typewriter", "anim_in_dur": 0.6,
"anim_out": null, "loop_anim": null,
"keyframes": [{"prop": "alpha", "at": 0, "value": 0}]}]},
{"type": "sticker", "name": "st", "segments": [
{"resource_id": "<capcut sticker id>", "start": 1, "duration": 2}]},
{"type": "effect", "name": "fx", "segments": [
{"effect": "Betamax", "start": 0, "duration": 2, "params": [60],
"character": false}]},
{"type": "filter", "name": "grade", "segments": [
{"filter": "BW_2", "start": 0, "duration": 7, "intensity": 45}]}
],
"srt": {"file": "subs.srt", "track": "subs", "offset": 0,
"size": 7.5, "color": "FFFFFF", "bold": true, "align": 1}
}
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 160 lines · 71 tokens per session scan A 0f0e367fc86e
capcut-draft is a skill published in the GitHub repository Belagum/promptcut (1 stars, last pushed 7d ago), licensed MIT. It adds 71 tokens to every session and 2,099 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
demo-assembly
Assembling and rendering a demo video with Remotion and ffmpeg — the timeline, transitions, captions, zoom-to-click, audio ducking, and diagnosing a render that comes out wrong. Use when rendering the final video, customising its look, or fixing timing, black frames or missing audio.
capcut-edit
Edit CapCut / JianYing video projects — read and write subtitles, timing, speed, volume, templates, animations (fade/ken-burns), and cut long-form to shorts. Use when the user mentions capcut, jianying, subtitles, video editing, draftcontent.json, draftinfo.json, or cutting videos.
compose-creative
Core creative engine — turns a calendar post into finished images or video using 4 modes (anchor-compose, enhance-extend, style-referenced, pure-creative) with brand assets, AI generation, and human approval before credits are spent. Triggers on "/compose-creative", "generate the creative", "make the image for P07"…
c2pa-sign
Embed C2PA provenance manifests in AI-generated assets — the machine-readable AI disclosure EU AI Act Article 50 expects (enforceable since 2 Aug 2026), plus the human-approval record for the Article 50(4) exemption. Triggers on "/c2pa-sign", "sign this asset", "provenance", "content credentials", "AI disclosure"…
setup
One-time setup for image and video generation APIs — Vertex AI, WaveSpeed, and Higgsfield credentials stored persistently and verified before use. Triggers on "/setup", "set up socialforge", "configure the APIs", "add credentials", "connect vertex", "credential status", or before any creative production on a new…
full-pipeline
Run the complete production pipeline end to end — parse the calendar, match assets, compose creative, adapt copy, build previews, run reviews, finalize delivery — with checkpoints between phases. Triggers on "/full-pipeline", "run the whole month", "produce the month", "end to end", "do everything", "full production…