SureForge
SkillAn instruction-only skill that makes agents research before they ask, plan before they build, and verify before they call anything done.
- workflow
- verification
- quality
- skill
32 medical-writing and research Agent Skills built for Chinese clinicians and graduate students: think in Chinese, write in English or Chinese, search PubMed, apply supervisor feedback and handle peer-review revisions — the full path from topic to revision (public beta, PolyForm NC 1.0.0)
dyz-med-paper-skills is duyunzhoudyz's source-available medical-writing skill collection (339 stars as of writing, 2026-11-10), designed around how Chinese clinicians and graduate students actually work: think the research through in Chinese, then write the manuscript in English or Chinese; pick topics from real ward cases and thesis requirements; connect Chinese and English literature with the writing; land supervisor comments as paragraph-level edits and revision checklists; balance graduation requirements, journal budgets and institutional wording; and resume from a record when work is interrupted. There are 32 dyz-med-* skills organized as Agent Skills SKILL.md directories, with dyz-med-orchestrator as the entry point — when you don't know where to start, it reads your paper folder and names the next problem and the missing materials. The repo ships teaching examples, a full skill index (docs/skills.md) and a bilingual docs site at https://duyunzhoudyz.github.io/dyz-med-paper-skills/. The installer is offline and uses only the Python standard library: it copies all 32 skills plus license files, prints PASS with the skill count and path, and stops when a same-name skill differs. Claude Code and Codex both have explicit paths (.claude/skills/ and .agents/skills/).
For Chinese clinicians and medical graduate students whose time is carved up by wards and rotations: each stage — topic, literature, Chinese/English drafts, supervisor revisions, submission and rebuttal — has a matching skill, and research groups can adopt it as a shared paper workflow.
git clone https://github.com/duyunzhoudyz/dyz-med-paper-skills.git cd dyz-med-paper-skills # Claude Code (personal skills dir): python scripts/install_codex.py --target "~/.claude/skills" python scripts/install_codex.py --target "~/.claude/skills" --check # Codex (defaults to ~/.agents/skills/): python scripts/install_codex.py python scripts/install_codex.py --check # then open your paper folder in the agent and describe the task; # dyz-med-orchestrator is the entry point when you don't know where to start
Public beta: continuous-writing quality, per-host behavior and an independent quality assessment of medical content are not fully evaluated, and writing quality depends on model, materials and task instructions. Online retrieval currently covers PubMed only — no CNKI account is bundled, and multi-database systematic reviews need other sources. Local scripts do not make the session offline: model services may receive session content, so follow your institution's rules and never submit identifiable patient data or hospital credentials. Built-in Word export suits basic review drafts; complex typesetting, equations and tracked changes need other tools. Environment: Python 3.10+ per the README. Licensed PolyForm Noncommercial 1.0.0 — commercial use requires separate written authorization (COMMERCIAL-LICENSE.md); third-party materials remain under their own terms. Research judgment and the final manuscript still need review by you and relevant professionals; passing checks is not publication. 339 stars as of writing (2026-11-10).
Repository
duyunzhoudyz/dyz-med-paper-skillsCharted
An instruction-only skill that makes agents research before they ask, plan before they build, and verify before they call anything done.
Guizang's agent skill turns your repo's recent updates into a finished promo video — real components, code-rendered animation, an original code-written soundtrack — no video editor involved.
An agent skill that geolocates a photo with nothing readable in it — OSM geometry, elevation skylines and satellite tiles turn “where was this taken?” into a measured answer with an error radius.