bioinformatics-agent-skills
AGENTS.md67 starsChanged 19 days agoMITView on GitHub
本文件为在本仓库中工作的 AI 编码助手提供上下文与约定。 bioinformatics-agent-skills 是一个 Agent Skills 技能库,为计算生物学领域提供 可被 AI 助手自动发现与触发的标准化技能包。每个技能是一个独立目录,含 SKILL.md、 scripts/、references/。 必须走这个流程,不要手工造目录: 然后填写真实的 SKILL.md 与脚本,最后校验: 完整规范见 docs/SKILL_SPEC.md。 新增技能后,必须同步更新 packages/biopipe/registry.py 的 REGISTRY, 否则 biopipe <name> 命令不可用,CI 的集成测试会失败。 这是科研工具,代码必须*真实可运行*:
In this repository17 other files it gives its agents.
- alphafold-dbSkill
- bio-database-fetchSkill
- bioproject-searchSkill
- bulk-rnaseqSkill
- ena-downloadSkill
- literature-searchSkill
- ml-peptide-generationSkill
- molecular-dynamicsSkill
- paper-fulltext-extractSkill
- peptide-designSkill
- protein-docking-hdockSkill
- protein-docking-lightdockSkill
- protein-interactionsSkill
- protein-structureSkill
- pymol-structure-visualizationSkill
- scientific-figuresSkill
- single-cell-rnaseqSkill
The file
Files
AGENTS.md
本文件为在本仓库中工作的 AI 编码助手提供上下文与约定。
这是什么项目
bioinformatics-agent-skills 是一个 Agent Skills 技能库,为计算生物学领域提供
可被 AI 助手自动发现与触发的标准化技能包。每个技能是一个独立目录,含 SKILL.md、
scripts/、references/。
目录约定
skills/<skill-name>/ 技能包,一个技能一个目录,kebab-case 命名
packages/biopipe_core/ 共享工具层(配置/日志/HTTP/CLI/IO)
packages/biopipe/ 统一命令行入口与技能注册表
scripts/ 仓库级工具(校验器、骨架生成器)
tests/ pytest 测试
docs/ 文档与技能目录开发约定
1. 新增技能
必须走这个流程,不要手工造目录:
python scripts/init_skill.py <name> --summary "..." --data-desc "..."然后填写真实的 SKILL.md 与脚本,最后校验:
python scripts/validate_skills.py --skill <name> --strict2. 脚本规范
- 共享能力一律从
biopipe_core导入,不要重复实现路径解析、日志、HTTP 重试 - 禁止硬编码绝对路径;输入输出路径来自
argparse - 禁止裸
print(),用get_logger("biopipe.<skill>") - 入口统一用
bootstrap_cli(description, skill_name, main) - 重量级依赖(scanpy/OpenMM/MDAnalysis 等)必须优雅降级,核心算法内置纯 Python 实现
- 每个脚本都要能被
--help正常响应,且不依赖真实数据
3. SKILL.md 规范
name必须与目录名一致,kebab-casedescription第三人称,含触发语句,中英文关键词都要有- 必须包含 Overview / When to Use / Quick Start / Workflow / Parameters / Outputs / Validation / References 八个章节
完整规范见 docs/SKILL_SPEC.md。
4. 注册新技能
新增技能后,必须同步更新 packages/biopipe/registry.py 的 REGISTRY,
否则 biopipe <name> 命令不可用,CI 的集成测试会失败。
提交前自检
python -m pytest tests -q # 测试
python scripts/validate_skills.py --strict # 规范校验
python -m compileall -q skills packages scripts # 语法三项全绿再提交。
科研正确性底线
这是科研工具,代码必须真实可运行:
- 不要写
pass # TODO冒充实现 - 不要返回硬编码的假数据
- 算法可以简化,但要说明简化了什么、以及何时不适用
- 结果必须能被
SKILL.md的 Validation 章节检查
提交信息规范
feat: add <skill-name> for <purpose>
fix: <what was broken>
docs: <what changed>
refactor: <what moved>
ci: <what changed>Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

