#671 · Primary category: AI Agents & Automation

COG-second-brain

agent-loop agent-plugins agent-skills agentic ai-agents claude-code closed-loop cursor garry-tan gbrain gstack knowledge-management loop-engineering obsidian people-crm productivity second-brain self-evolving verification-harness worker-agents

Self-evolving second brain with 33 AI skills, 10 agents, and a verification harness that prevents self-grading, working across Claude Code, Cursor, Kiro, Gemini CLI, and Codex.

Project last updated:08/25/26

GitHub Stars

1.1K

Forks

131

Contributors

6

License

MIT

Why we included this project

This repo drops a ready-made library of 33 skills and 10 agents into an existing coding assistant like Claude Code, Cursor, Kiro, Gemini CLI, or anything else that reads markdown, turning plain .md files into a structured knowledge base without standing up a database. What sets it apart is the closed-loop design: worker agents produce the notes and decisions, while separate read-only verifier agents check the output, so the agent that does the work never grades its own homework. That guardrail matters if you've watched autonomous agents quietly reinforce their own mistakes. The bundled skills cover everyday productivity like braindumps and daily briefs, product work such as user stories, PRDs, and release notes, plus a people CRM, and teams already living in Obsidian, Git, and issue trackers will find it a practical way to keep agent output checked and memory tidy.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category