Novelty Check
Checks whether a research idea has already been published, using literature search and a second-model review.
- Stars
- 18,482
- Licence
- MIT
- Files
- 1 file
What it does
When you describe a research idea, this skill tells the agent to break the idea into its core claims, search arXiv, Google Scholar, Semantic Scholar and recent conference papers for overlapping work, and read the closest abstracts. It then asks a separate Claude reviewer (through a local claude-review MCP bridge) to judge novelty and produces a structured report with a score, a PROCEED / PROCEED WITH CAUTION / ABANDON recommendation, and a one-sentence statement of what is new. It also stores a trace of the search and review in a project folder.
Great for
- Researchers checking whether an idea is worth a pilot experiment before investing months in it
- Students who want a structured first pass on prior work before writing a proposal
- Codex users who want a Claude model to act as an independent reviewer
Skip it if
- Final novelty verdicts for publication: the skill itself says it is only an early gate, and it can miss papers or misjudge them, so a human literature review is still needed
- Projects that cannot send search queries or review content to external services, because it relies on web search and an external reviewer bridge
- Non-research tasks, since it is designed for judging academic novelty of methods
Try asking
- “查新: check whether this sparse attention method for long-context LLMs has already been published”
- “Check novelty of my idea: using diffusion models to repair corrupted audio tracks”
- “有没有人做过 applying retrieval augmentation to protein structure prediction?”
- “Before I implement it, verify this training-data selection strategy is novel against 2025-2026 arXiv papers”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
novelty-check.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip novelty-check.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip novelty-check.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Give a specific method description with the mechanism and the problem it solves, since vague ideas produce vague searches
- Read the closest-prior-work table yourself and check the cited papers, because the reviewer output is model-generated
- Install this after the base skills/skills-codex package if you use it as a Codex override, as the file instructs
What's inside
- SKILL.md
Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.