paper-plan

Compiles paper outlines from a claim graph with evidence maps, figure plans, and citation plans.

Updated Jul 15, 2026
One-click install
npx skills add https://github.com/1549171506/shizhengmubangai1 --skill paper-plan-1549171506
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: paper-plan
Source: https://github.com/1549171506/shizhengmubangai1/tree/main/Desktop/123/empiricalwiki/.claude/skills/paper-plan
Command: npx skills add https://github.com/1549171506/shizhengmubangai1 --skill paper-plan-1549171506

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning validated research claims and experiment results into a coherent, venue-compliant paper outline is slow and error-prone. This Skill compiles a claim graph from a research wiki into a complete paper plan, ensuring every section is justified by a claim and every claim is covered by a section. ## Core Features & Use Cases - Claim-driven outline generation: Builds an evidence map linking claims to experiments and papers, then derives section structure, page budgets, and narrative arc for venues like ICLR, NeurIPS, ICML, ACL, CVPR, and IEEE. - Figure, table, and citation planning: Produces a figure/table plan sourced from experiment results and pre-fetches BibTeX from DBLP, CrossRef, and Semantic Scholar, flagging unconfirmed citations. - Mandatory LLM review: Submits the outline to a review LLM acting as an area chair to test narrative persuasiveness and evidence sufficiency before writing begins. - Use Case: After running experiments that support three claims in your research wiki, invoke the Skill with those claim slugs and --venue ICLR to receive a PAPER_PLAN.md with section outlines, page budgets, citation coverage, and review feedback. ## Quick Start Ask the assistant to run /paper-plan with your supported claim slugs and a target venue, for example: plan an ICLR paper from claims primary-claim and ablation-claim with the title "Scaling Widget Networks".

Frequently Asked Questions about paper-plan

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

FAQPage Schema
How do I generate a paper outline from research claims?

Invoke the Skill with the slugs of supported or weakly supported claims plus a target venue flag such as --venue ICLR. It compiles an evidence map from the wiki, derives a section outline with page budgets, and writes a PAPER_PLAN.md to wiki/outputs/.

What venues does the paper planning workflow support?

The --venue flag accepts ICLR, NeurIPS, ICML, ACL, CVPR, and IEEE. The venue determines page limits and formatting constraints used when allocating the section page budget.

Can I plan a paper without experiment results?

No. The Skill requires at least one experiment evidence entry backing the target claims; otherwise it errors and suggests running experiment design and execution first. Purely proposed claims trigger warnings or errors.

How are citations verified in the citation plan?

Each referenced wiki paper's BibTeX is pre-fetched from DBLP, then CrossRef, then Semantic Scholar as fallback. Entries that fail all sources are marked [UNCONFIRMED] with manual check URLs, and a coverage report summarizes verified counts.

What happens if the review LLM is unavailable?

The workflow degrades to a Claude self-review of the outline and labels the report as single-model review, noting that cross-model verification was unavailable. The outline is still produced and written to the wiki.