Verified:GitHub

draft / research

Generate evidence-backed research briefs and learning-loop artifacts from stage trails.

Updated Aug 27, 2026

Search Tags:#agentic workflows #dataset generation #research drafting #learning-loop variants #context snapshotting #rejection analytics #calibration reporting

Ready to run

npx skills add https://github.com/Collectively-Gary/GaryOS --skill draft-research

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: draft / research
Source: https://github.com/Collectively-Gary/GaryOS/tree/main/data/skills/draft/research
Command: npx skills add https://github.com/Collectively-Gary/GaryOS --skill draft-research

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Drafting reliable research from an action plan while preserving the system’s audit trail and correctly routing learning-loop variants.

Core Features & Use Cases

Plan-driven brief generation: Loads the latest stage plan, then produces a research brief that satisfies every “Truths” item in the Success criteria.
Context ref-aware drafting: Walks context references, resolves merges, and snapshots relevant fields into the trail when decisions depend on context state.
Learning-loop variants: Produces variant-specific artifacts for per-action learning loops, including:
- generic research (default) that writes a brief artifact for the action folder,
- rejections that aggregates rejected review patterns into a report,
- dataset that emits JSONL training rows plus a strict manifest,
- calibration that builds a quadrant-style join from gate recommendation vs operator answer vs verify outcomes.

Quick Start

Run the draft/research skill for a playbook research action with learning_variant set to dataset to emit dataset.jsonl and dataset.manifest.yaml for the specified learning window.

Frequently Asked Questions about draft / research

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

FAQPage Schema

How do I generate an evidence-backed research brief from an action plan?▼

To generate an evidence-backed research brief, the skill loads the latest stage plan and produces a document satisfying every Success criteria item. It walks context references, resolves merges, and snapshots relevant fields into the trail when decisions depend on context state.

How do I create a JSONL dataset and manifest from learning-loop trails?▼

Creating a JSONL dataset requires setting the learning_variant to dataset for the playbook research action. The skill processes the specified learning window, emits dataset.jsonl training rows, and writes a strict dataset.manifest.yaml file.

What is a calibration quadrant table in learning-loop analytics?▼

A calibration quadrant table is a variant-specific artifact that joins gate recommendations, operator answers, and verify outcomes. It is generated when the learning_variant is set to calibration, providing a matrix to analyze learning-loop performance.

How does context snapshotting work during research drafting?▼

Context snapshotting works by walking context references, resolving merges, and recording needed context fields into the trail. This ensures that when drafting decisions depend on context state, the exact state is preserved for auditability.

What happens when the skill encounters malformed windows or unreadable trail files?▼

When encountering malformed windows or unreadable trail files, the skill handles them safely without crashing the workflow. It enforces observability of Success criteria and records standardized trail entries while managing these edge cases gracefully.

Can I aggregate rejected review patterns into a report for learning loops?▼

Yes, you can aggregate rejected review patterns by setting the learning_variant to rejections. The skill processes the action's trail context and produces a dedicated report analyzing the rejected patterns within the specified learning window.