video-spec-generate

Generate validated spec.yaml files for ace-web video programs from source content.

1|2|Updated Apr 1, 2026
One-click install
npx skills add https://github.com/dimagi-internal/ace --skill video-spec-generate
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: video-spec-generate
Source: https://github.com/dimagi-internal/ace/tree/main/skills/video-spec-generate
Command: npx skills add https://github.com/dimagi-internal/ace --skill video-spec-generate

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manually creating video specs for ace-web is time-consuming, requires maintaining separate per-template prompts, and often leads to inconsistent formatting, missing placeholders, or specs that fail TTS validation. This skill eliminates that manual work by providing a single, template-agnostic workflow that auto-populates, validates, and posts complete video specs from any source input.

Core Features & Use Cases

  • Template-Agnostic Generation: Works with every ace-web video template without requiring new prompts or skill updates when templates are added to the platform.
  • Multi-Source Ingestion: Accepts input as Connect program page URLs, pasted text briefs, or structured prospect research artifacts to auto-fill spec placeholders.
  • Built-In Validation: Enforces Connect brand voice rules, TTS-compatible word budgets per narration beat, and strict stat grounding from source data only to prevent fabricated claims.
  • Real-World Use Case: You have a new Connect campaign program page and need a 60-second overview video spec. Use this skill to auto-extract program details, fill the template skeleton, validate the narration fits audio timing requirements, and post the completed spec directly to your ace-web workspace without manual formatting.

Quick Start

Use the video-spec-generate skill to turn the provided Connect program page URL into a complete, validated video spec for the 60s-campaign-overview template and post it to the default dimagi-team workspace.

Frequently Asked Questions about video-spec-generate

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

FAQPage Schema
How do I generate video specs from a program page URL automatically?

To generate video specs from a program page URL, this skill ingests the page content, auto-populates a template-agnostic spec.yaml file, and validates it against TTS word budgets. It accepts Connect program page URLs, pasted text briefs, or structured prospect research artifacts as input to eliminate manual authoring.

Can I use pasted text briefs to create TTS-compatible video specs for ace-web?

Yes, you can use pasted text briefs to create TTS-compatible video specs for ace-web. The system accepts unstructured or structured text briefs, automatically enforcing TTS-compatible word budgets per narration beat and Connect brand voice rules before posting the completed spec to your target workspace.

Does this video spec generation workflow support all ace-web templates without custom prompts?

Yes, this video spec generation workflow supports all ace-web templates without requiring custom prompts. It uses a template-agnostic approach to populate placeholders, meaning you do not need to update or maintain separate prompts when new video templates are added to the platform.

What is the best way to prevent fabricated statistics in automated video spec files?

The best way to prevent fabricated statistics in automated video spec files is to use strict source grounding. This skill enforces stat grounding from the provided source data only, ensuring all claims in the generated spec.yaml are validated against the original Connect program page or pasted research artifacts.

How do I auto-post completed video specs to an ace-web workspace?

To auto-post completed video specs to an ace-web workspace, the skill uses authenticated API calls to push the validated spec.yaml files directly to your target environment. This eliminates manual file handling and ensures the posted specs have already passed brand voice and TTS validation.

Can I generate multi-angle narration variants for partnership-focused video templates?

Yes, you can generate multi-angle narration variants for partnership-focused video templates. The skill supports creating multiple narration angles within a single template-agnostic generation pass, while still enforcing strict TTS-compatible word budgets per beat for each variant.