Analyze quality signals, optimize prompts, and generate a learning report.
npx skills add https://github.com/knsiuss/orion --skill self-improve-knsiuss
Please help me install this Agent Skill. Skill: self-improve Source: https://github.com/knsiuss/orion/tree/main/workspace/skills/self-improve Command: npx skills add https://github.com/knsiuss/orion --skill self-improve-knsiuss
This Skill automates the process of improving EDITH's performance by analyzing past interactions, optimizing its internal prompts, and identifying potential new skills.
Instruct EDITH to run its self-improvement cycle.
High-intent search queries and answers about installing and using this skill.
This Skill automates prompt optimization by analyzing past interaction quality signals to identify struggle areas and propose targeted prompt improvements. It requires your explicit approval before applying any mutable prompt changes to the system.
Identifying recurring workflows for skill creation involves analyzing past user interactions to detect repeated patterns. The system drafts these detected workflows as candidate auto-skills, presenting a learning report for your review and approval.
You run an automated learning cycle by manually triggering the self-improvement process. This cycle reviews performance feedback signals to identify areas where the AI struggles, subsequently proposing prompt optimizations and generating a summary learning report.
Yes, the AI self-improvement process requires explicit user approval for prompt changes. It also requires user review of candidate auto-skills drafted from detected recurring workflows before any new skill creation or system modification occurs.
You should manually trigger an AI self-improvement cycle after a period of extensive system use. This ensures the system continuously learns from accumulated interaction data, optimizing its responses and task execution capabilities over time.
curiositechLogs skill invocations and scores output quality from usage data.
1J6K21Review code changes across correctness, readability, architecture, security, and performance.
agdevGuide creation of Claude Code agents with architecture and prompt engineering.
QcasaresDesign and optimize AI prompts using CoT, few-shot, and XML techniques.
