skill-graph
Skill1 starsChanged 7 months agoMITView on GitHub
小能核心技能 — 扫描所有Skill,推断并维护Skill之间的关系图谱(前置/后续/依赖/互补/替代/包含),输出Mermaid可视化图+JSON关系表。触发词:技能图谱/Skill关系/依赖关系/调用链/关系图谱
2 other projects use a version of this.
Adopted by
Most recently taken up by
- bryanwhl/skill-graphtheir own version, first seen May 6, 2026
- Vix0007/vixero-skillstheir own version, first seen Apr 24, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- bryanwhl/skill-graph★ 2same name
- Vix0007/vixero-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- 小能Skill
- agent-capability-mapSkill
- skill-auditSkill
- skill-card-genSkill
- skill-forgeSkill
- skill-registry-scanSkill
The file
Files
Skill 图谱维护
v2.0 | 升级:标准化Mermaid输出+具体推断规则+质量评估
触发词
- "技能图谱" / "Skill 关系" / "依赖关系" / "调用链" / "关系图谱"
关系类型
| 类型 | 符号 | 含义 | 判断规则 |
| 前置 | ← | 调用A之前应先完成B | A的输入=B的输出,或A的SKILL.md明确说"先调用B" |
| 后续 | → | A完成后通常调用B | A的输出是B的输入,或A的SKILL.md说"结果交给B" |
| 依赖 | ⟵ | A必须有B才能运行 | A的工具链/脚本/API依赖B提供的数据或工具 |
| 互补 | ↔ | A和B一起用效果更好 | 同Agent下、同项目中经常一起调用 |
| 替代 | ≈ | A和B解决同一问题 | description关键词高度重叠(>70%),但方法不同 |
| 包含 | ⊃ | A包含B作为子流程 | A的skills/目录下有B,或A的流程中明确调用B |
流程
Step 1:全量扫描 registry
读取 memory/skill-registry.json,获取所有 status=active 的 Skill。
Step 2:推断关系(三类信号)
信号一:声明关系(SKILL.md中明确提到,最高优先级)
- 搜索关键词:"与XX的关系"、"协作"、"调用"、"依赖"、"前置"、"后续"、"交给"、"先做"
- 命中 → 标注
source: declared
信号二:结构关系(从目录结构推断)
- A的skills/目录下有B →
contains - 同一Agent下的Skill → 标注
same_agent: true - A的references/引用了B的名称 →
depends_on
信号三:语义关系(从description推断)
- 计算description关键词重叠率
- 重叠>70% →
alternative - 重叠30-70%且同Agent →
complementary - 重叠<30% → 无关系
Step 3:写入 registry
在每个Skill条目中添加 relations 字段:
{
"relations": [
{"type": "prerequisite", "target": "skill-id", "source": "declared"},
{"type": "contains", "target": "sub-skill-id", "source": "structural"}
]
}Step 4:输出标准关系图
Mermaid格式(可直接渲染):
graph TD
A[skill-forge] -->|前置| B[skill-audit]
B -->|后续| C[skill-card-gen]
D[小能] -->|包含| A
D -->|包含| B
D -->|包含| CJSON关系表:
{
"nodes": [
{"id": "skill-forge", "agent": "小能", "type": "meta"},
{"id": "skill-audit", "agent": "小能", "type": "management"}
],
"edges": [
{"from": "skill-forge", "to": "skill-audit", "type": "subsequent", "source": "declared"}
],
"stats": {
"total_skills": 6,
"with_relations": 6,
"coverage_rate": 1.0,
"declared_relations": 4,
"inferred_relations": 3
}
}Step 5:输出关系摘要
📊 Skill 图谱 · {日期}
━━━━━━━━━━━━━━━━━━━━
节点:{N}个 Skill | 边:{N}条关系 | 覆盖率 {X}%
🔗 关键调用链:
skill-forge → skill-audit → skill-card-gen → agent-capability-map
skill-registry-scan → (所有其他技能的数据源)
📁 包含关系:
小能 ⊃ {skill-audit, skill-registry-scan, skill-card-gen, skill-graph, agent-capability-map, skill-forge}
⚠️ 孤立节点:{无任何关系的Skill}
💡 建议:{补建关系或考虑合并}质量标准
- 关系覆盖率 = 有关系的Skill / 总Skill,目标≥80%
- 声明关系占比 = 声明关系 / 总关系,目标≥50%(推断关系不可靠)
- 孤立节点 = 0为最佳
输出文件
memory/skill-registry.json— 更新relations字段memory/skill-graph-{date}.md— 本次图谱快照
注意事项
- 只记录有明确依据的关系,不臆造
- 推断的关系标注
source: inferred,声明的标注source: declared - 每次增量更新,不重写整个图谱
- Mermaid图中的节点ID必须与registry.json的name字段一致
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.

