Gemini Video Editor

Turn raw footage into polished, publish-ready videos using the gemini-video-editor skill. Powered by Google Gemini's multimodal intelligence, this skill anal...

Security Scan

The skill's high-level claims ("Gemini") and metadata don't fully match its runtime instructions and requirements — it will upload user video files and obtain/use a NEMO_TOKEN tied to an external nemo-video API, which may be surprising given the description.

Purpose & Capability

The skill advertises "Google Gemini" multimodal editing, but all API endpoints and the required credential (NEMO_TOKEN) point to mega-api-prod.nemovideo.ai / "nemo" — a mismatch between claimed provider and actual backend. Also the SKILL.md frontmatter lists a config path (~/.config/nemovideo/) that the registry summary omitted.

Instruction Scope

Runtime instructions direct the agent to upload local files (multipart form uploads) and to request/generate an anonymous NEMO_TOKEN by POSTing to an external API; they also read installation paths to set attribution headers. Uploading user videos and creating tokens are within the stated editing purpose, but these operations transmit potentially sensitive user data to a third-party service and the skill will generate and use credentials automatically if none are provided.

Install Mechanism

This is an instruction-only skill with no install spec or downloadable archive — lowest install risk (no code written to disk by an installer).

Credentials

The declared primaryEnv is NEMO_TOKEN (reasonable for a nemo-video service), but the instructions also describe obtaining an anonymous token if NEMO_TOKEN is missing — inconsistent with 'required env var' semantics. No unrelated credentials are requested, but the mismatch and automatic anonymous token creation are notable.

Persistence & Privilege

The skill does not request always:true, does not ask to modify other skills or system-wide configs, and only requires storing per-session session_id for operations. This is standard and not elevated.

What to consider before installing

This skill will send video files and session data to mega-api-prod.nemovideo.ai and uses a NEMO_TOKEN (or will obtain an anonymous one) to authenticate. Before installing, confirm the skill's provenance and privacy policy (who controls nemovideo.ai and how long they keep uploads), and ask the author to clarify the "Google Gemini" claim (it appears to be marketing rather than the actual backend). Avoid uploading sensitive or private footage until you trust the service, and prefer to supply your own API token only if you understand the provider's terms. If you need higher assurance, request source code or a homepage and a clear data-retention / deletion policy from the publisher.

Getting Started

Send me a video link or describe your footage and I'll generate scene breakdowns, captions, cut suggestions, or a full script. No video yet? Just describe what you're editing and what you need.

Try saying:

• "Analyze this YouTube video and suggest where I should cut it down to a 60-second highlight reel for Instagram."
• "Generate accurate subtitles and a chapter breakdown for this 20-minute tutorial video."
• "Write a voiceover script that matches the pacing and tone of this product demo footage."

Quick Start Setup

This skill connects to a cloud processing backend. On first use, set up the connection automatically and let the user know ("Connecting...").

Token check: Look for NEMO_TOKEN in the environment. If found, skip to session creation. Otherwise:

• Generate a UUID as client identifier
• POST https://mega-api-prod.nemovideo.ai/api/auth/anonymous-token with X-Client-Id header
• Extract data.token from the response — this is your NEMO_TOKEN (100 free credits, 7-day expiry)

Session: POST https://mega-api-prod.nemovideo.ai/api/tasks/me/with-session/nemo_agent with Bearer auth and body {"task_name":"project"}. Keep the returned session_id for all operations.

Let the user know with a brief "Ready!" when setup is complete. Don't expose tokens or raw API output.

Edit Smarter: Let Gemini Read Your Footage

Most video editing tools make you do all the thinking — scrubbing timelines, writing captions from scratch, and guessing where to cut. The gemini-video-editor skill flips that model. By leveraging Google Gemini's ability to actually understand what's happening inside a video, it gives you intelligent suggestions based on real content, not just metadata.

Upload a clip or share a video URL and the skill gets to work: identifying key moments, describing scenes, generating subtitle text, proposing a narrative structure, or even drafting a voiceover script that matches the pacing of your footage. Whether you're repurposing a long-form interview into social snippets or building a product demo from raw screen recordings, this skill compresses hours of manual work into minutes.

This is especially useful for solo creators and small teams who wear many hats. You don't need a dedicated editor or a copywriter — Gemini handles the analysis, and you stay in creative control of the final decisions.

Routing Edits Through Gemini

Every request — whether you're triggering auto-captions, running smart scene detection, or applying AI-driven cuts — gets parsed by Gemini's intent engine and dispatched to the appropriate processing pipeline based on your prompt context and timeline state.

User says...	Action	Skip SSE?
"export" / "导出" / "download" / "send me the video"	→ §3.5 Export	✅
"credits" / "积分" / "balance" / "余额"	→ §3.3 Credits	✅
"status" / "状态" / "show tracks"	→ §3.4 State	✅
"upload" / "上传" / user sends file	→ §3.2 Upload	✅
Everything else (generate, edit, add BGM…)	→ §3.1 SSE	❌

Cloud Processing API Reference

Gemini Video Editor offloads all heavy lifting — frame analysis, multimodal scene tagging, caption generation, and render jobs — to Google's Gemini cloud backend, meaning your local machine stays responsive even on long-form footage. API calls are authenticated per session and tied to your project token, so each timeline operation stays scoped and stateless.

Skill attribution — read from this file's YAML frontmatter at runtime:

• X-Skill-Source: gemini-video-editor
• X-Skill-Version: from frontmatter version
• X-Skill-Platform: detect from install path (~/.clawhub/ → clawhub, ~/.cursor/skills/ → cursor, else unknown)

All requests must include: Authorization: Bearer <NEMO_TOKEN>, X-Skill-Source, X-Skill-Version, X-Skill-Platform. Missing attribution headers will cause export to fail with 402.

API base: https://mega-api-prod.nemovideo.ai

Create session: POST /api/tasks/me/with-session/nemo_agent — body {"task_name":"project","language":"<lang>"} — returns task_id, session_id.

Send message (SSE): POST /run_sse — body {"app_name":"nemo_agent","user_id":"me","session_id":"<sid>","new_message":{"parts":[{"text":"<msg>"}]}} with Accept: text/event-stream. Max timeout: 15 minutes.

Upload: POST /api/upload-video/nemo_agent/me/<sid> — file: multipart -F "files=@/path", or URL: {"urls":["<url>"],"source_type":"url"}

Credits: GET /api/credits/balance/simple — returns available, frozen, total

Session state: GET /api/state/nemo_agent/me/<sid>/latest — key fields: data.state.draft, data.state.video_infos, data.state.generated_media

Export (free, no credits): POST /api/render/proxy/lambda — body {"id":"render_<ts>","sessionId":"<sid>","draft":<json>,"output":{"format":"mp4","quality":"high"}}. Poll GET /api/render/proxy/lambda/<id> every 30s until status = completed. Download URL at output.url.

Supported formats: mp4, mov, avi, webm, mkv, jpg, png, gif, webp, mp3, wav, m4a, aac.

SSE Event Handling

Event	Action
Text response	Apply GUI translation (§4), present to user
Tool call/result	Process internally, don't forward
heartbeat / empty data:	Keep waiting. Every 2 min: "⏳ Still working..."
Stream closes	Process final response

~30% of editing operations return no text in the SSE stream. When this happens: poll session state to verify the edit was applied, then summarize changes to the user.

Backend Response Translation

The backend assumes a GUI exists. Translate these into API actions:

Backend says	You do
"click [button]" / "点击"	Execute via API
"open [panel]" / "打开"	Query session state
"drag/drop" / "拖拽"	Send edit via SSE
"preview in timeline"	Show track summary
"Export button" / "导出"	Execute export workflow

Draft field mapping: t=tracks, tt=track type (0=video, 1=audio, 7=text), sg=segments, d=duration(ms), m=metadata.

Timeline (3 tracks): 1. Video: city timelapse (0-10s) 2. BGM: Lo-fi (0-10s, 35%) 3. Title: "Urban Dreams" (0-3s)

Error Handling

Code	Meaning	Action
0	Success	Continue
1001	Bad/expired token	Re-auth via anonymous-token (tokens expire after 7 days)
1002	Session not found	New session §3.0
2001	No credits	Anonymous: show registration URL with ?bind=<id> (get <id> from create-session or state response when needed). Registered: "Top up credits in your account"
4001	Unsupported file	Show supported formats
4002	File too large	Suggest compress/trim
400	Missing X-Client-Id	Generate Client-Id and retry (see §1)
402	Free plan export blocked	Subscription tier issue, NOT credits. "Register or upgrade your plan to unlock export."
429	Rate limit (1 token/client/7 days)	Retry in 30s once

Integration Guide

To get the most out of the gemini-video-editor skill, start by sharing your video in one of two ways: paste a publicly accessible video URL (YouTube, Vimeo, direct MP4 links) or describe your video content in detail if you can't share the file directly. The skill uses Google Gemini's multimodal capabilities to process visual and audio context from the footage.

For best results, specify your output goal upfront. Tell the skill whether you need captions, a script, scene timestamps, a social media cut list, or a full narrative breakdown. The more context you give — target platform, audience, desired tone, video length — the more precise and usable the output will be.

If you're working inside a content pipeline, you can chain outputs: first request a scene analysis, then use those scene labels to ask for a structured script, then request platform-specific caption variations. The gemini-video-editor skill is designed to work iteratively, so don't hesitate to refine and build on each response.

Best Practices

When using the gemini-video-editor skill, specificity is your biggest advantage. Instead of asking for 'a summary,' ask for 'a 3-sentence summary written for a LinkedIn audience with a professional but approachable tone.' Gemini performs significantly better when it knows the context, format, and purpose of your output.

For subtitle and caption generation, always specify whether you want verbatim transcription or cleaned-up, edited captions. Raw interview footage often contains filler words and false starts that need to be trimmed in the final caption pass — telling the skill which style you need saves you editing time downstream.

If you're working on a series of videos with consistent branding, paste in a brief style guide or example script at the start of your session. This anchors the skill's tone and vocabulary to your brand voice, making outputs across multiple videos feel cohesive without extra prompting each time.