physarum
Skill9 starsChanged 3 months agoMITView on GitHub
Slime-mold collective intelligence. Multi-pulse parallel exploration where agents interconnect through the Board to form one brain. Use when a problem needs breadth-first probing with natural convergence through shared memory.
In this repository29 other files it gives its agents.
- AGENTS.mdAGENTS.md
- analyze-failuresSkill
- anti-debugging-techniquesSkill
- anti-drift-disciplineSkill
- binary-protection-bypassSkill
- code-obfuscation-deobfuscationSkill
- distillSkill
- dwarf-expertSkill
- emulation-reSkill
- find-skillsSkill
- ida-mcpSkill
- code-reviewSkill
- diagnosing-bugsSkill
- domain-modelingSkill
- implementSkill
- prototypeSkill
- researchSkill
- resolving-merge-conflictsSkill
- tddSkill
- to-specSkill
- writing-great-skillsSkill
- mobile-dynamic-reSkill
- native-static-reSkill
- re-escalationSkill
- re-routerSkill
- stormSkill
- symbolic-execution-toolsSkill
- vm-and-bytecode-reverseSkill
- web-reverse-engineeringSkill
The file
Files
Physarum — 黏菌集体智能
何时触发
- 问题空间需要多角度探索
- 单一视角不够，需要集体智慧互联
- 需要发散后自然收敛（而非人工裁判）
与 Storm 的区别
| Storm | Physarum | |
| 模式 | 对抗辩论 | 协同探索 |
| 角色关系 | advocate vs critic | 同一大脑的多个神经元 |
| 收敛方式 | 裁判综合 | 群体自然聚焦 |
| 适用场景 | 判断题审计 | 探索性问题解决 |
如何使用
- 启用：
/physarum on <model1>,<model2> - 打开 topic：
board action=open topic="my-problem" goal="..." - 启动探索：
spawn action=physarum topic="my-problem" question="..." angles=["angle1", "angle2", ...] - 查看结果：
board action=read topic="my-problem" - 收尾：
board action=close→distill
运行机制
- Pulse 1（发散）：N 个触角并行探索不同角度
- Pulse 2..N-1（构建）：每个触角读取 Board 上他人发现，互相构建、深化、交叉
- 最后一轮（综合）：所有触角综合集体发现为统一答案
- 工作过程封闭在子 topic 内，结束后主 topic 仅新增 1 条综合结论（零副作用）
不要在以下情况使用
- 需要对抗性审计（用 /storm）
- 简单直接的实现任务（直接 spawn action=run）
- 只有一个明确方向不需要探索
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.