Self Improving Agent

Most agents repeat the same mistakes because they're stateless. This skill gives you patterns for agents that actually learn: feedback loops that capture what worked and what failed, memory curation that turns session logs into actionable rules, and a promotion system that graduates proven patterns from MEMORY.md into enforced rules in CLAUDE.md. It's built as five sub-skills (Remember, Extract, Promote, Review, Status) that form an improvement cycle, plus scripts for pattern extraction and regression detection. Use this when you're building agents that run repeatedly and you're tired of watching them make the same errors. The core insight is solid: auto-memory captures everything, but curation is what makes it useful.

skills.sh installs: 241
GitHub stars: 1.1K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add borghei/claude-skills --skill self-improving-agent --agent claude-code

Installs into .claude/skills of the current project.

Files

Self-Improving Agent

The agent captures what went wrong or right in a session, curates it in memory, and proposes promoting a lesson into a permanent rule only when it has recurred and a person has approved it. Rules compound: a wrong rule repeats its mistake in every later session, so promotion is a reviewed change, never an automatic one.

Core Capabilities

Memory curation — a layered memory stack (CLAUDE.md → MEMORY.md → session), review protocol, and promotion criteria for graduating learnings into enforced rules.
Feedback loops — outcome classification, signal extraction, and a capture template that turn every task result into a structured learning.
Regression detection — metrics, thresholds, and a response protocol that flags performance degradation within a few sessions.
Skill extraction — criteria and a 4-step process to graduate proven patterns into standalone skill packages.
Meta-learning — adaptive capture strategy and anti-pattern detection so the agent learns what is worth learning.
Continuous calibration — confidence scoring and belief revision for resolving contradictions across learned knowledge.