v0.2.0 — expanded skills, agents & rules
Claude Code infrastructure for academic research — skills, agents, hooks, and rules for LaTeX, bibliography, proofreading, research design, and project management workflows.
Highlights since v0.1.0
Coverage: 61 skills · 15 agents · 18 auto-loaded rules · 9 hooks.
New agents: code-review, blindspot, claim-verify, code-paper-auditor, fatal-error-check, gemini-research, codex-research, artifact-coherence-auditor, reproducibility-auditor.
New rules: doi-verification, mark-unverified, no-hardcoded-results, python-uv, severity-gradient, latex-hygiene, paper-code-consistency, subagent-write-guard, spec-before-quality, audit-before-fix.
New skill families: bibliography (bib-parse, bib-coverage, bib-filter), LaTeX scaffolding (latex-scaffold, latex-template), document formats (docx, pdf, xlsx), research design (causal-design, experiment-design, synthetic-data), writing voice (voice-analyzer, voice-editor), and paper-book companions.
Packaging fix: corrected the npm package name and repository URL to flonat-research so npx flonat-research resolves correctly.
Install
npx flonat-researchOr clone for full customisation:
git clone https://github.com/flonat/claude-research.git
cd claude-research
./scripts/setup.sh