ppt-research-pack

To extract claims and evidence for slide-ready content, this skill synthesizes uploaded reports and external retrieval into structured claims, evidence points, and pageworthy chunks. It assigns stable IDs to ensure downstream slide assembly uses verifiable inputs.

You need structured research packs when your task-pack metadata identifies content gaps before slide creation. This skill runs conditionally if research_required is true, preventing slides from proceeding with missing or unverified information.

To track uncertainties when researching slide content, this skill cross-validates external retrievals and annotates conflicting evidence in a dedicated risks_or_uncertainties section. It avoids inventing facts and ensures all uncertain data points are explicitly flagged.

Yes, you can generate a structured research pack automatically from uploaded documents and targeted external research. The skill ingests these inputs and outputs a formatted research-pack.json file containing all claims, evidence, and chunks with stable IDs.

Slide-deck workflow research requires stable IDs like claim_id, evidence_id, and chunk_id to maintain reference integrity. These stable identifiers allow downstream modules like ppt-info-pack and ppt-storyboard to accurately reference specific content blocks.

A limitation of automating research pack generation is that it relies entirely on task-pack.json configuration and will not run unless research_required is set to true. Additionally, it avoids inventing facts, so outputs are constrained by the quality of uploaded reports and external retrievals.