sci-paper

v0.38.0

Scientific paper writing, source-traced review, and de-AI rewriting for top-tier journals (ApJ/MNRAS/PRD/JCAP-class) and NSF/NIH proposals. Ships 12 skills and 40 tools under one typed standard: scientific-integrity blockers that cannot be waived as style preferences, a zero-target L0 lexicon, ranked advisories, explicit measurement states so missing calibration never reads as clean, claim-first rewriting gated on hard scientific-fidelity eligibility, a mechanical length budget with a machine-built removal map and a required-shrink gate, corpus-calibrated salience-hierarchy, domain-register (with an exhaustive zero-hit vocabulary audit), collocation, and discourse-texture (given/new cohesion, epistemic hedging) axes, each judged against the field's own per-section distribution at the unit it was calibrated at, a residue axis that gates the trace an edit leaves, A-R review with narrative-spine and adversarial-escalation dimensions, registry-verified bibliographies (CrossRef, DataCite, arXiv), pixel-measured figure review, and isolated multi-reviewer orchestration. Learned scores are field-similarity triage, never an authorship verdict. See CHANGELOG.md for per-version history.

Claude Code2 Skills

By skymanbpLicense: MIT5 GitHub starsUpdated last week

Directory evidence

Runtimes: Claude Code
Parsed components: 2 skill or MCP entries
Source updated: Sep 16, 2026
Manifest status: Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install sci-paper for Claude Code

Installs for the current user

claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install sci-paper@agent-plugin-marketplace

Paste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually

$ git clone https://github.com/skymanbp/sci-paper

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The repository root is the plugin root.

Plugin files

sci-paper/

├── .claude-plugin/plugin.json
├── skills/paper/SKILL.md
└── skills/proposal-polish/SKILL.md

Included Skills2

paperskills/paper/SKILL.md

Load the unified SCIPAPER_STANDARD writing guidance, formula derivation conventions, scientific-integrity rules, canonical L0 examples, and field-specific reference anchors for ApJ/MNRAS/PRD-level papers. Use when writing or reviewing scientific content.

proposal-polishskills/proposal-polish/SKILL.md

Funding-proposal editing mode (NSF Project Summary/Description, NIH Specific Aims, fellowship proposals). Keeps the vision-and-feasibility register a paper would trim, enforces claim-feasibility matching, edits the score-forming first pages hardest, and applies the shared L0 lexical policy. Never invents preliminary data, partners, funding history, or letters, and is not for evading AI-use disclosure.

Plugin manifests1

.claude-plugin/plugin.json

{
  "name": "sci-paper",
  "version": "0.38.0",
  "description": "Scientific paper writing, source-traced review, and de-AI rewriting for top-tier journals (ApJ/MNRAS/PRD/JCAP-class) and NSF/NIH proposals. Ships 12 skills and 40 tools under one typed standard: scientific-integrity blockers that cannot be waived as style preferences, a zero-target L0 lexicon, ranked advisories, explicit measurement states so missing calibration never reads as clean, claim-first rewriting gated on hard scientific-fidelity eligibility, a mechanical length budget with a machine-built removal map and a required-shrink gate, corpus-calibrated salience-hierarchy, domain-register (with an exhaustive zero-hit vocabulary audit), collocation, and discourse-texture (given/new cohesion, epistemic hedging) axes, each judged against the field's own per-section distribution at the unit it was calibrated at, a residue axis that gates the trace an edit leaves, A-R review with narrative-spine and adversarial-escalation dimensions, registry-verified bibliographies (CrossRef, DataCite, arXiv), pixel-measured figure review, and isolated multi-reviewer orchestration. Learned scores are field-similarity triage, never an authorship verdict. See CHANGELOG.md for per-version history.",
  "author": {
    "name": "skymanbp",
    "email": "[email protected]"
  },
  "homepage": "https://github.com/skymanbp/sci-paper",
  "repository": "https://github.com/skymanbp/sci-paper",
  "license": "MIT",
  "keywords": [
    "claude-code",
    "claude-code-plugin",
    "agent-skills",
    "scientific-writing",
    "academic-writing",
    "technical-writing",
    "paper-review",
    "peer-review",
    "manuscript",
    "de-ai",
    "latex",
    "arxiv",
    "astrophysics",
    "cosmology",
    "research-tools",
    "writing-assistant",
    "reproducibility",
    "style-distillation",
    "research-ideation",
    "brainstorm"
  ]
}