Skip to content
velonx*SKILLS

Academic Researchv1.2.0

Find, read and summarise peer-reviewed papers and build a short cited literature review. Use when the user asks what the research says, wants papers on a topic, needs a literature review, or asks to evaluate a study's methods.

Author
Velonx
License
MIT
Updated
Sep 25, 2026

Overview

Search scholarly sources, screen papers by relevance and quality, and synthesise them into a literature review with proper citations and an honest read of the evidence.

When to Use

  • "What does the research say about…", "find papers on…", "literature review of…"
  • Evaluating a specific study ("is this paper any good?").
  • Preparing background for a thesis, grant or technical decision.

Installation

coming soonVelonx CLI:velonx install academic-research
  1. Get the skill folder

    git clone --depth 1 https://github.com/velonx/agent-skills.git
    The skill is agent-skills/skills/academic-research/. Or open it on GitHub and download the files.
  2. Give SKILL.md to your agent as instructions

    Add it to the system prompt, custom instructions or rules file your agent reads. Keep the rest of the folder next to it if the skill refers to examples or scripts.
  3. Ask for the task in your own words

    The agent follows the skill when your request matches it. You can also name it: “use academic-research”.

Agents move these settings between versions — if a path doesn’t match, check your agent’s docs.

Usage

  1. Scope. Define the question, the time window and inclusion criteria (study type, population, field). Confirm with the user if unclear.
  2. Search scholarly indexes (Google Scholar, Semantic Scholar, PubMed, arXiv, SSRN — whichever fit the field). Use 2–3 query variants plus the key authors' names once you find them.
  3. Screen titles and abstracts. Keep 5–15 papers. Prefer systematic reviews and meta-analyses, then large or well-controlled studies.
  4. Read each kept paper's abstract, methods and results. For each record: design, sample size, main finding, effect size, limitations, funding/conflicts.
  5. Synthesise by theme, not paper by paper. Note where findings agree, conflict, or where evidence is thin.
  6. Output:
    • One-paragraph summary of the state of evidence.
    • Themed sections with in-text citations (Author, Year).
    • A table: paper · design · n · finding · quality note.
    • Reference list in the style the user asks for (default APA).
  7. Mark preprints as not peer-reviewed.

Examples

Prompt: "Summarise the research on spaced repetition for learning programming."

The agent finds reviews on the spacing effect, narrows to studies in computing education, builds a table of 8 papers, and concludes that evidence is strong for factual recall but limited for problem-solving skills.

Requirements & compatibility

Your agent needs

  • Web access — search and open pages

Written for

GenericClaudeOpenAIGeminiCursor

Limitations

  • Full text may be paywalled; the agent says when it only read the abstract.
  • Never invents citations — if a reference can't be found again, it is dropped.
  • Not a substitute for a formal systematic review.

Files

Changelog

  • 1.2.0 — Preprints are now flagged explicitly; added quality-note column.
  • 1.1.0 — Themed synthesis instead of per-paper summaries.
  • 1.0.0 — Initial release.

Contributing

This skill lives in velonx/agent-skills. Improvements go through a pull request, reviewed by maintainers.