shortform-production

Produce vertical short-form videos with hook previews, editorial review, and verified Metricool API scheduling.

3.6k|688|Updated Mar 28, 2026
Ready to run
npx skills add https://github.com/ericosiu/ai-marketing-skills --skill shortform-production-ericosiu
💬 Prompt for ChatGPT / Claude Chat▼
Please help me install this Agent Skill.
Skill: shortform-production
Source: https://github.com/ericosiu/ai-marketing-skills/tree/main/shortform-production
Command: npx skills add https://github.com/ericosiu/ai-marketing-skills --skill shortform-production-ericosiu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests, and includes scripts (resource) and references (resource) components.

What problem does it solve? Short-form video production often lacks consistent editorial standards, evidence-based visuals, and safe publishing workflows, leading to unreadable captions, unsupported claims, and duplicate or misconfigured social posts. ## Core Features & Use Cases - Hook Experimentation: Generate three four-second opening treatments (problem, evidence, process) for one clip, preview them at phone size, and select the clearest before rendering the full master. - Editorial Quality Evaluation: Score finished edits against a 10-dimension weighted rubric with hard gates, score caps, and a JSON scorecard schema covering hook structure, visual proof, captions, and technical finish. - Verified Metricool Delivery: Schedule Instagram Reels through a Python API client with dry-run mode, account verification, duplicate prevention, media hash verification, and receipt-based idempotency. - Use Case: A creator has a recorded talking-head video and wants one polished Instagram Reel. The skill guides the edit using the V5 style preset, produces three opening previews, scores the final cut with the eval rubric, then schedules it through Metricool only after explicit authorization. ## Quick Start Use $shortform-production on this recording to make three four-second opening previews, choose the clearest, and finish one captioned short with the source CTA without publishing.

Frequently Asked Questions about shortform-production

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I schedule Instagram Reels with the Metricool API?▼

Create a private profile JSON with blog_id, user_id, instagram handle, brand label, and timezone, then prepare a manifest with video path, caption, CTA, and future time. Run scripts/metricool.py in dry-run mode first, then add --apply within explicit authorization to schedule with verification.

How do I test different hooks for a short-form video?▼

Write three treatment cards for one clip covering problem, evidence, and process openings, each with exact timestamps and a hypothesis. Render only four-second previews, review them at 360 px width with sound, and select the clearest before producing the full master.

Does the Metricool script make network calls by default?▼

No, the script runs in dry-run mode by default and makes no network calls. Adding the --apply flag executes the schedule or move operation, and the client verifies the account against /admin/simpleProfiles before any write.

What Python version and dependencies are required?▼

The API client requires Python 3.10 or later and the requests library version 2.32 or higher, installed via python3 -m pip install -r requirements.txt. The METRICOOL_API_TOKEN must be set as an environment variable, never stored in files.

Why does scheduling fail with a duplicate or occupied slot error?▼

The client rejects scheduling when a matching caption already exists in the queue, the Instagram slot is occupied at the same minute, or a verified receipt differs from the current post. Inspect the queue and reconcile saved receipts before retrying.

Can the skill publish videos automatically without approval?▼

No, installation does not authorize publishing, new accounts, automations, or public releases. Scheduling through the Metricool API only occurs within explicit task authorization, and the skill reports status only with matching delivery evidence.