Scaffold agent-ready academic research repositories with SOTA files, source ledgers, wiki memory, MCP setup, and project-local skills.
SaferSkills independently audited create-academic-research (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Scaffold a four-entity academic research project with one command: SOTA → survey → contributions → papers, one root bibliography, MCP-verified citations, agent skills included.
npm create academic-research@latest my-projectmy-project/
├── README.md AGENTS.md CLAUDE.md
├── references.bib # THE single bibliography (1:1:1 with the SOTA)
├── .mcp.json # arxiv + semantic-scholar + dblp always on
├── pyproject.toml # uv workspace: one venv for all contributions
├── Makefile # make check | pdfs | survey | contribution | paper
├── scripts/check.py # structure validator
├── sota/ # papers/<citekey>/{paper.pdf,synthesis.md,metadata.yaml} + index + queue
├── survey/ # survey.tex + committed survey.pdf + coverage.md
├── contributions/ # self-contained badge-compliant units (_template/)
└── papers/ # one folder per venue submission (_template/)The wizard asks for a title, a one-line topic, optional MCP servers (openalex pre-checked; zotero and overleaf opt-in), and installs the companion academic-research-skills project-locally. Flags: --yes, --no-install-skills, --no-git.
standard synthesis + metadata with citation graph), its bib entry, and its index row. A citation exists only if a scholarly MCP lookup produced it — provenance is recorded per paper and enforced by make check.
synthesis, groups the SOTA, and ends with gaps & research directions. coverage.md makes updates diff-driven when the SOTA changes.
with a report detailed enough to write papers from.
template, packaged artifacts, correspondence, and immutable submission archives.
Generated projects need: git, make, python3 (≥3.11), latexmk, and uv (uvx runs the MCP servers). No Node at runtime (except the optional openalex server).
npm install
npm run typecheck
npm test
npm pack --dry-runThe test suite needs python3 ≥ 3.11 on PATH (one test validates generated TOML with tomllib); CI uses 3.12.
Tag-driven. Bump package.json, commit, tag vX.Y.Z, push the tag; the release workflow validates, publishes to npm, and creates a GitHub release.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.