Capture errors and learnings into structured .learnings markdown files.
npx skills add https://github.com/syahravi/openclaw --skill self-improvement-syahravi
Please help me install this Agent Skill. Skill: self-improvement Source: https://github.com/syahravi/openclaw/tree/main/skills/self-improvement Command: npx skills add https://github.com/syahravi/openclaw --skill self-improvement-syahravi
Captures and organizes learnings, corrections, errors, and feature requests so teams and agents can continuously improve and avoid repeating the same mistakes. It provides a lightweight, consistent format for logging incidents and promoting broadly applicable insights into project memory.
Use the self-improvement skill to immediately log a correction or error to .learnings/LEARNINGS.md or .learnings/ERRORS.md after a failure or user correction.
High-intent search queries and answers about installing and using this skill.
To log errors and learnings for AI agent workflows, you can append structured markdown entries to .learnings/ERRORS.md and LEARNINGS.md after command failures or user corrections. This preserves context and suggested fixes for continuous improvement.
Preventing repeat mistakes involves capturing corrections and API failures in structured markdown files, then promoting high-value insights into project memory files like CLAUDE.md or AGENTS.md so agents can reference past learnings.
Track feature requests by appending standardized entries to .learnings/FEATURE_REQUESTS.md when a missing capability is requested. This logs the context and metadata needed to address the gap later.
Yes, you can automate error detection and reminders by integrating optional hook scripts and workspace injection points. These trigger automatic logging and reminders when commands or external APIs fail during agent execution.
To promote recurring agent learnings into project memory, you move high-value entries from .learnings/ files into project-level configuration files like CLAUDE.md, AGENTS.md, or SOUL.md, making best practices discoverable across the workspace.
g-lokOrganize learnings, errors, and feature requests into persistent markdown logs.
1012LoninLog learnings, errors, and corrections to .learnings with standardized metadata.
FelipeMainesInject contextual rules into prompts via an 8-layer pipeline.
QcasaresCapture and process software development learnings and errors for knowledge retention.
