About this skill

Problem

jarvis-research-longtext-v1 addresses information loss in paper reading: a short summary can miss the Architecture, equation derivations, and ablation details, while a generic model summary may skip paragraphs or oversimplify math. It targets algorithm and computer-science papers on Arxiv and aims to convert Abstract, Intro, Methodology, and Experiments into a professional, verifiable long-form article.

Workflow

The skill uses three phases:
- Atomic indexing: it tries to download the PDF first, retries on failure, and asks for a file if Arxiv retrieval fails, then indexes paragraphs across sections to reduce omission.
- Template-driven writing: it produces a Title, TLDR, Past Problems, Proposed Method, Key Benefits, Model Architecture, and Experiments section, with inline source notes such as [Ref: Section 3.2].
- Reflection loop: it checks template coverage, paragraph coverage, completeness of equations like Eq. 1, and the input-to-output logic, then rewrites weak sections.

Boundaries

It is not intended for one-sentence TLDR requests or non-algorithm papers. Ambiguous formulas or diagrams are marked [needs verification] rather than inferred.

Screenshots

Use Cases

Before a tech talk, turn an Arxiv algorithm paper into a detailed speaker-ready long-form article.
While reproducing a search paper, verify coverage of inputs, core modules, outputs, and numbered equations.
During algorithm research, convert multiple Arxiv papers into structured notes on problems, methods, and benefits.
When reviewing paper comprehension, check whether the long article covers derivations, ablations, and sources.

Best For

Algorithm researchers preparing close-reading reports who need complete architecture, equations, and experiment details.
Engineers doing technical research who need Arxiv papers broken into problems, methods, benefits, and model flow.
Instructors writing technical sharing notes who need a clear, checkable, source-linked long article.
Supervisors reviewing paper comprehension who need to detect missing formulas or ablation analysis.

Related Skills

Scholar Reasoning Distiller

@user_231765ee/django-plus

Extracts an eight-layer reasoning fingerprint from scholar texts and applies a nine-module constraint to run peer review, supervision, lecturing, or panel discussion with quantitative scores and qualitative comments.

Knowledge Management 2026.08.30 ↓ 608