continuous-learning
基于 instinct 的持续学习系统,通过 hooks 观察会话,创建带置信度评分的 atomic instincts, 并将高置信度 instinct 演进为 skills/commands/agents。v2.1 增加项目级 instincts 防止跨项目污染。
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- continuous-learning
- origin
- adapted from ECC
- version
- 2.1.0
- description
- 基于 instinct 的持续学习系统,通过 hooks 观察会话,创建带置信度评分的 atomic instincts, 并将高置信度 instinct 演进为 skills/commands/agents。v2.1 增加项目级 instincts 防止跨项目污染。
02 资源文件 10
- 脚本agents/observer-loop.sh7.6 KB
- 参考文档agents/observer.md7.2 KB
- 参考文档agents/openai.yaml279 B
- 脚本agents/session-guardian.sh6.2 KB
- 脚本agents/start-observer.sh7.3 KB
- 参考文档config.json135 B
- 脚本hooks/observe.sh15.1 KB
- 脚本scripts/detect-project.sh7.6 KB
- 脚本scripts/instinct-cli.py58.4 KB
- 脚本scripts/test_parse_instinct.py32.3 KB
03 SKILL.md 正文
Continuous Learning v2.1 - Instinct-Based Architecture
一个高级学习系统,通过 atomic "instincts" 将 Claude Code 会话转化为可重用知识。
v2.1 增加 项目级 instincts — React 模式保留在 React 项目中,Python 约定保留在 Python 项目中,通用模式(如"始终验证输入")全局共享。
何时激活
- 设置从 Claude Code 会话自动学习
- 配置基于 hooks 的行为提取
- 调优学习行为的置信度阈值
- 审查、导出或导入 instinct 库
- 将 instincts 演进为完整 skills、commands 或 agents
- 管理项目级 vs 全局 instincts
- 将 instincts 从项目级提升到全局级
Instinct 模型
一个 instinct 是一个小的学习行为:
---
id: prefer-component-colocation
trigger: "当多个组件使用相同状态时"
confidence: 0.7
domain: "frontend-architecture"
source: "session-observation"
scope: project
project_id: "a1b2c3d4e5f6"
project_name: "points-frontend"
---
# 组件状态共置
## 行为
当多个组件使用相同状态时,考虑将状态提升到共同祖先。
## 证据
- 在 5 个实例中观察到状态提升模式
- 用户在 2026-03-29 将组件本地状态改为共享状态
属性:
- Atomic -- 一个 trigger,一个 action
- Confidence-weighted -- 0.3 = 试探性,0.9 = 几乎确定
- Domain-tagged -- code-style, testing, git, debugging, workflow 等
- Evidence-backed -- 跟踪创建它的观察
- Scope-aware --
project(默认)或global
存储结构
~/.claude/homunculus/
+-- projects.json # 项目注册表: hash -> name/path
+-- observations.jsonl # 全局观察(fallback)
+-- instincts/
| +-- personal/ # 全局自动学习的 instincts
| +-- inherited/ # 全局导入的 instincts
+-- evolved/
| +-- agents/ # 全局生成的 agents
| +-- skills/ # 全局生成的 skills
| +-- commands/ # 全局生成的 commands
+-- projects/
+-- a1b2c3d4e5f6/ # 项目 hash(来自 git remote URL)
+-- project.json # 项目元数据
+-- observations.jsonl
+-- instincts/
| +-- personal/ # 项目特定自动学习
| +-- inherited/ # 项目特定导入
+-- evolved/
+-- skills/
+-- commands/
+-- agents/
项目检测
系统自动检测当前项目:
CLAUDE_PROJECT_DIRenv var(最高优先级)git remote get-url origin-- hash 生成项目 IDgit rev-parse --show-toplevel-- fallback- 全局 fallback -- 如果未检测到项目,instincts 进入全局范围
置信度评分
| 分数 | 含义 | 行为 |
|---|---|---|
| 0.3 | 试探性 | 建议但不强制 |
| 0.5 | 中等 | 相关时应用 |
| 0.7 | 强 | 自动批准应用 |
| 0.9 | 几乎确定 | 核心行为 |
置信度增加当:
- 模式被重复观察
- 用户不纠正建议的行为
- 相似 instincts 同意
置信度减少当:
- 用户明确纠正行为
- 模式长时间未观察到
- 出现矛盾证据
与 Skills 的区别
| 特性 | Skills | Instincts |
|---|---|---|
| 粒度 | 完整工作流 | Atomic 行为 |
| 触发 | 手动调用 | 自动观察 |
| 置信度 | 无 | 0.3-0.9 |
| 演化 | 直接成为 skill | 先 instinct 再 cluster |
相关工具
scripts/lib/memory_store.py- 底层存储接口- Error Experience Library - 错误模式的专门 instinct
命令接入
/instinct-status- 显示学习的 instincts/evolve- 将 instincts 聚类为 skills/commands/instinct-export- 导出 instincts 到文件/instinct-import- 从文件导入 instincts/promote- 将项目 instincts 提升到全局
最佳实践
- 观察 -- 启用 hooks 捕获所有工具调用
- 分析 -- 定期运行 observer agent 分析观察
- 演进 -- 将高置信度 instincts 演进为 skills
- 分享 -- 通过导出/导入在团队中共享 instincts
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.