Forgedemy
CatalogPreview source

charon-fan

Self-Improving Agent

Universal self-improvement system that learns from every skill interaction using multi-memory architecture.

Price

Free

Files

0

Rating

0.0

Reviews

0

Source

Source repo

Get freePreview source

About

Implements a feedback loop using semantic, episodic, and working memory layers to continuously evolve skill guidance based on real usage patterns. Auto-triggers on skill completion and errors via hooks, performs self-correction when guidance errors are detected, and can submit improvements as PRs. Based on 2025 lifelong learning research (SimpleMem, Evo-Memory) and designed to avoid infinite recursion during error handling.

By charon-fan

Identity GitHub charon-fan

What the agent sees

name

skills-sh-charon-fan-agent-playbook-self-improving-agent

description

Universal self-improvement system that learns from every skill interaction using multi-memory architecture.

Tags

self-improvementmemoryagentslifelong-learninghooksautomationTools: skills-cli, external-adapter, upstream-install

Technical details

Source repoOriginal GitHub linkPublisher site

Source repo

https://github.com/charon-fan/agent-playbook

Original GitHub link

https://github.com/charon-fan/agent-playbook

Publisher site

https://skills.sh/charon-fan/agent-playbook/self-improving-agent

Packaging note

Imported from the public skills.sh trending snapshot fetched at 2026-03-18T00:58:16.450Z. Snapshot rank #121 with 11500 weekly installs. Bundle files are not mirrored into Forgedemy.

Recent reviews

No reviews yet.