Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Video Production
v1.0.0Complete A/B video pipeline — storyboard, Veo 3 batch generation, browser preview with feedback loop, and ffmpeg assembly into final videos. Use when creatin...
⭐ 1· 634·11 current·11 all-time
by@omerflo
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description match the code: scripts call Veo (Google GenAI) and DALL·E, build previews, accept feedback, and assemble with ffmpeg — that capability set is coherent with a "video production" skill. However the registry metadata claims no required environment variables while the code and SKILL.md clearly require GOOGLE_API_KEY (and dalle_generator.py requires OPENAI_API_KEY). That metadata omission is an inconsistency worth flagging.
Instruction Scope
Runtime instructions and scripts operate on storyboard.json, feedback.json, and project files (clips/, assets/, timing.json) — all expected. They perform network calls to Google GenAI and OpenAI image endpoints to download generated assets, and call ffmpeg/ffprobe locally. The SKILL.md also instructs adding a cron for a quota watcher that "texts Master"; you should inspect scripts/quota_watcher.sh and any notification code to confirm where credentials or webhook targets are stored and whether it posts to external endpoints. The flow references a helper called "Muffin" (an assistant that suggests prompts) — understand where that runs and what data it receives.
Install Mechanism
No install spec in the registry; SKILL.md recommends creating a Python venv and pip installing google-genai, Pillow, requests. This is a low-risk, conventional setup (no random binary downloads or extract-from-URL).
Credentials
The skill requires API keys (GOOGLE_API_KEY for Veo/GenAI and OPENAI_API_KEY for DALL·E) and the code checks for those. The registry metadata advertised 'no required env vars', which is incorrect and lowers transparency. Also the SKILL.md suggests writing GOOGLE_API_KEY to ~/.zshenv; ensure you understand and trust the remote APIs before storing keys. No other unrelated secrets are requested in the code.
Persistence & Privilege
The skill does not request always:true and will not auto-install itself system-wide. It writes/reads files within its project directory and backs up replaced clips (e.g., .prev.mp4) — expected behavior for a content pipeline. It will call local utilities (ffmpeg, ffprobe, open).
What to consider before installing
What to check before installing/use:
- Do not trust the registry's 'no env vars' claim: the code requires GOOGLE_API_KEY (and some modules/scripts expect OPENAI_API_KEY). Only provide keys you control and understand the billing implications. Store them in a safe place and avoid using highly privileged or organization-wide keys.
- Inspect scripts/quota_watcher.sh and any notification/integration code (the SKILL.md references texting a "Master" and an Orchestrator that mentions Telegram/iMessage). Confirm it does not post your API keys or project files to external, unknown endpoints or require additional credentials for messaging.
- Run the skill in an isolated environment (dedicated venv, non-privileged account) and review network traffic if possible — the generators download assets from Google/OpenAI and the code writes files to the skill workspace.
- Expect the scripts to create/overwrite files in the project dir (clips/, outputs). Back up anything important first.
- Verify ffmpeg is installed from a trusted source and that the ffmpeg commands used match your needs.
- If you plan to grant this skill to an autonomous agent, be cautious: it can make network API calls and generate/overwrite files. The metadata inconsistency (env vars omitted) reduces transparency; ask the publisher to fix registry metadata and provide a short security note about notification endpoints before using broadly.
Confidence note: assessment is medium because the code appears to implement the described pipeline (not obviously malicious), but the metadata omission and notification/cron references are unexplained and warrant manual inspection.Like a lobster shell, security has layers — review code before you run it.
latestvk971tneqpa72vbbkey8askt2bd81rjcd
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
