Media-Editor-MCP: Skill for Claude Code

.claude/skills/pp-create-reel-once/SKILL.md

pp-create-reel-once is a skill for Claude Code from nguyenph88/Media-Editor-MCP. It costs 145 tokens per session (1,657 once invoked), scanned A, original, MIT.

A command for making a music-synchronised video reel in Adobe Premiere Pro where each source clip is used once. The reel ends naturally when the clips run out instead of repeating footage to reach a fixed length.

In plain words
What is it for?
Use `/pp-create-reel-once` to choose chronological or smart ordering and set how many musical downbeat intervals each clip should occupy.
Why use it?
It avoids repeated clips and fixed-duration padding when you have a small set of footage or want each clip’s full moment to appear.

Skill for Claude Code

Written for Claude Code: installed under .claude/.

This is nguyenph88/Media-Editor-MCP's own configuration. It tells Claude Code how to work on Media-Editor-MCP itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything Media-Editor-MCP configures →

Reuse

Borrowing it

Nothing to install: this file belongs to nguyenph88/Media-Editor-MCP. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/nguyenph88/Media-Editor-MCP/main/.claude/skills/pp-create-reel-once/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/nguyenph88/Media-Editor-MCP

Made for: Claude Code.

Wrote 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.

agentmods badge for pp-create-reel-once

README.md
[![agentmods](https://agentmods.dev/badge/skills/nguyenph88/media-editor-mcp/pp-create-reel-once/github.svg)](https://agentmods.dev/skills/nguyenph88/media-editor-mcp/pp-create-reel-once)
Your own site
<a href="https://agentmods.dev/skills/nguyenph88/media-editor-mcp/pp-create-reel-once"><img src="https://agentmods.dev/badge/skills/nguyenph88/media-editor-mcp/pp-create-reel-once/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for pp-create-reel-once

Your own site · 80×15
<a href="https://agentmods.dev/skills/nguyenph88/media-editor-mcp/pp-create-reel-once"><img src="https://agentmods.dev/badge/skills/nguyenph88/media-editor-mcp/pp-create-reel-once.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 145 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,657 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00145 $0.01657
Opus 5.5 $0.00058 $0.00663
Sonnet 5.5 $0.00029 $0.00331
Haiku 4.5 $0.00015 $0.00166

Measured 2mo ago against content hash bd700125dcf9, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

pp-create-reel-once 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 2mo 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.

.claude/skills/pp-create-reel-once/SKILL.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Create reel (each clip once) — natural-length beat edit

A variant of [[pp-create-reel]] for when you DON'T want to repeat footage to fill 59s. Each clip plays once; the reel ends when the clips run out (e.g. 5 clips ≈ 10-25s depending on spans). Cuts still land on downbeats. Verified tooling is identical — only the slot-allocation logic differs.

Inputs: music, footage folder, plus two knobs:

  • order: chronological (clips in folder/given order — story order) or smart (open on the highest-scoring clip as a hook, rest by motion). Default chronological for this mode (it's usually a "show my clips in order" use case).
  • beatsPerClip: how many downbeat-intervals each clip spans — 1, 2, 3, or auto (default). auto gives longer/richer source clips more beats (see Step 3). Spanning ≥2 beats gives a clip more screen time and suits cinematic / slow footage.

This mode does NOT force 59s, does NOT repeat clips, and has NO loop ending (each clip is unique — looping would require repeating the first clip). Everything else (vocal-skip, smart slices, dissolves, grade, punch-ins, title) works the same.

Pre-flight & footage

Same as [[pp-create-reel]] Step 2: premiere_health + analysis_health green; probe footage and reject any clip with an audio stream; one orientation. NO ~12-clips rule here — use exactly the clips the user has.

Step 1 — Analyze

  1. Vocal-skip (default on): transcribe → first vocal; music start S = first downbeat ≥ first_voice − 0.15 (instrumental → S=0).
  2. detect_beats → downbeats (source time). Index the downbeats at/after S as D = [D0, D1, …] (D0 = first downbeat ≥ S). bar = median(diff(D)).

Step 2 — Order the clips

  • chronological: keep the folder's natural sort / the order the user listed.
  • smart: run find_best_moments on each, put the single highest-scoring clip first (hook), then the rest by descending top-window motion.

Step 3 — Assign a beat-span to each clip

Per clip, decide span ∈ {1,2,3} = how many consecutive downbeat-intervals it occupies:

  • fixed (beatsPerClip=1|2|3): every clip gets that span.
  • manual (beatsPerClip=[3,1,2,…]): one span per clip, in order.
  • auto (default): a two-layer model — the SONG sets the cutting pace, then per-clip ACTIVITY distributes spans around it. This makes energetic tracks cut fast and cinematic tracks linger, while still giving the most dynamic clips the longer holds.
    1. Song layer — classify_song(music) → recommendedSpanCenter (energetic ≈ 1.3, balanced ≈ 1.8, cinematic ≈ 2.5) from tempo + onset density + percussive fraction. This is the average beats-per-clip.
    2. Clip layer — find_best_moments returns absolute, cross-clip-comparable activityMotion / activitySharpness (whole-clip means, before per-clip normalization). Min-max them across the clip set; activity = 0.7·motionN + 0.3·sharpN, giving each clip an activityRank ∈ [0,1].
    3. Combine: span = round(spanCenter + (activityRank − 0.5) × 1.5), clamped to [1,3]. (So on a cinematic song every clip leans 2-3 but the liveliest still top out at 3; on an energetic song most clips are 1 and only the most dynamic reach 2.)
    4. Cap by what the clip can fill: shrink span while clipDuration < span×bar + 0.3. A short clip can't span 3 even if lively; a long static clip stays low.
    • This replaces the old raw-duration rule (which wrongly gave long STATIC clips 3 beats). beatsPerClip (fixed/manual) still overrides everything.
  • A clip's slot duration = D[k+span] − D[k] (sum of its bars). Get its slice from find_best_moments(clip, window_seconds = slotDur + 0.1, count=1) → top window start. (In chronological mode this still picks the BEST window within the clip — order refers to clip sequence, not within-clip.)

Read the full file on GitHub · 55 lines

Changes

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.

  1. 2mo ago First seen · 55 lines · 0 tokens per session scan A bd700125dcf9

Subscribe to this mod's changes

pp-create-reel-once is a skill published in the GitHub repository nguyenph88/Media-Editor-MCP (1 stars, last pushed 5mo ago), licensed MIT. It adds 145 tokens to every session and 1,657 once invoked, about $0.0006 per session on Opus 5.5. 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.

Related

pp-create-reel-once compared with skills from other repositories

pp-create-reel-once next to 6 skills from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
pp-create-reel-oncenguyenph88/Media-Editor-MCPSkill15mo agoMITA145Claude Code
local-ai-agentsmicrosoft/ai-agents-for-beginnersSkill80ktodayMITA200Claude Code, Codex
imagegenopenai/codexSkill136ktodayApache-2.0A113Codex
playwright-climicrosoft/playwrightSkill100ktodayApache-2.0A19Claude Code
browser-automationopenclaw/openclawSkill395ktodayMITA36Claude Code
next-cache-components-adoptionvercel/next.jsSkill144ktodayMITA95Claude Code
chat-pet-sprite-creationmicrosoft/vscodeSkill196ktodayMITA53Claude Code, Codex