菜单
Systematically extract lessons from tasks, compress them into reusable rules, and store them in the appropriate memory tier (Local, Global, or Private). Use this skill at the end of every major mission or experiment to build your 'Second Brain'.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
API security best practices, OWASP API Security Top 10 2023, authentication, authorization, rate limiting, and API gateway security. Use when securing REST APIs, GraphQL, OAuth 2.0 implementations, or API gateway configurations.
Cloud security best practices for AWS, GCP, and Azure. Use when securing cloud infrastructure, configuring IAM, hardening S3 buckets, setting up VPC security, compliance (SOC2, ISO 27001), or auditing cloud resources for misconfigurations.
Container and Kubernetes security hardening, image scanning, runtime protection, and supply chain security. Use when securing Docker containers, Kubernetes clusters, scanning images for vulnerabilities, or implementing runtime security controls.
Central dispatcher for the kinhluan skills ecosystem. Use when the user says "dùng skill phù hợp", "use right skill", "kinhluan skills", or when multiple skills could apply. Detects workflow stage, artifact existence, and intent to route to the correct skill.
Offensive security, penetration testing, and vulnerability exploitation. Use when conducting security assessments, red team exercises, web application pentesting, network exploitation, or learning ethical hacking techniques. Covers PTES methodology, reconnaissance, exploitation, post-exploitation, and reporting.
Security threat modeling and risk assessment using STRIDE, Attack Trees, and DREAD scoring. Use when designing secure architecture, conducting security design reviews, assessing risks, or creating threat models for applications, systems, or cloud infrastructure.
| name | second-brain-reflection |
| description | Systematically extract lessons from tasks, compress them into reusable rules, and store them in the appropriate memory tier (Local, Global, or Private). Use this skill at the end of every major mission or experiment to build your 'Second Brain'. |
| metadata | {"tags":["knowledge-management","reflection","compression","second-brain","learning-loop"]} |
"He who learns but does not think is lost! He who thinks but does not learn is in great danger." — Confucius (adapted for Agents)
This skill transforms raw experience into compressed, actionable tri thức (knowledge). It ensures that every failure is a lesson and every success is a reusable pattern.
Trigger this protocol after completing a Mission, PR, or Experiment.
Evaluate the task through the lens of Sun Tzu's Five Factors:
why-strategic-rationale?Identify 1-3 critical insights. Ask:
Do NOT save conversational fluff. Use the Rule/Stratagem Format:
Condition/Context → Action/Constraint. (Reason).Pattern → Optimization. (Impact).Rule: Large JSON (>500MB) → Use orjson, not stdlib. (5x faster serialization).Agent must determine the scope before saving:
| Scope | Location | Action |
|---|---|---|
| Project Specific | ./GEMINI.md | Add to "Shared Rules" section. |
| Private/Task Finding | ~/.gemini/tmp/skills/memory/MEMORY.md | Add to local project index. |
| Universal Truth | ~/.gemini/GEMINI.md | Add to "Personal Preferences" section. |
After processing, the agent should present the knowledge in this format:
### 🧠 Second Brain Sync — [Task Name]
**1. Strategic Audit (Sun Tzu):**
- [Factor]: [Finding]
**2. Compressed Knowledge:**
- `Rule: [Context] → [Action]. ([Why]).`
**3. Saving Location:**
- [Path to file]
Second Brain Reflection is a closing ritual for every major workflow. Trigger automatically at the end of:
After completing each stage:
After completing each stage:
After each strategic assessment:
After major coding tasks:
Before marking any task as DONE in collaborative-engineering-agent Kanban, ask:
"What is the one rule from this task that would save time on the next similar task?"
If the answer is non-trivial, run Second Brain Reflection before closing the task.
art-of-war: Provides the Audit framework.master-framework: Lessons feed back into Layer 0 (WHY) and Layer 2 (DESIGN).collaborative-engineering-agent: Reflection Gate before marking tasks DONE.kinhluan-router: Reflection results update routing heuristics ("last time X happened, Y skill was most useful").