self-improving-agent
A universal self-improving system that learns from skill experiences by maintaining semantic, episodic, and working memory. It uses hooks to trigger pattern extraction, self-correction, and skill updates based on session outcomes.
Is self-improving-agent safe to install?
Review the source first: our audit of self-improving-agent's source files found 3 shell commands, 4 external URLs, file reads and writes (high risk). Every command and URL listed appears verbatim in the skill's source. The skill executes arbitrary bash commands via hooks and modifies local files including SKILL.md and memory directories.
How we audit skills: our security review methodology.
Who is this skill for?
Developers using AI agents who want to automate the refinement of agent workflows, documentation, and code patterns based on past performance.
What can you do with it?
- Extracting reusable patterns from completed tasks
- Self-correcting guidance after detecting errors or failed bash commands
- Updating skill documentation and checklists with validated improvements
- Tracking confidence levels for specific patterns across multiple sessions
How good is this skill?
Quality score: 8/10. The documentation provides a clear architecture and specific file paths for memory management. The hook integration strategy is well-defined, though the reliance on manual path replacement for shell scripts requires careful user configuration.
What does the skill file contain?
# Self-Improving Agent > "An AI agent that learns from every interaction, accumulating patterns and insights to continuously improve its own capabilities." — Based on 2025 lifelong learning research ## Overview This is a **universal self-improvement system** that learns from ALL skill experiences, not just PRDs. It implements a complete feedback loop with: - **Multi-Memory Architecture**: Semantic + Episodic + Working memory - **Self-Correction**: Detects and fixes skill guidance errors - **Self-Validation**: Periodically verifies skill accuracy - **Hooks Integration**: Auto-triggers on sk...
Frequently asked questions
How does the agent decide when to update a skill?
It promotes changes when the user requests an update, a pattern recurs across multiple episodes, a test proves current guidance is wrong, or the change is low-risk documentation.
What happens when a bash command fails?
The on_error hook triggers, capturing the error context from memory/working/last_error.json to initiate a self-correction workflow.
Where are the learned patterns stored?
Abstract patterns reside in memory/semantic-patterns.json, while specific experiences are stored in memory/episodic/.
Related skills
find-skills
2.3MUsers seeking to extend agent capabilities with specialized tools, workflows, or knowledge packages
The find-skills skill enables agents to search for, discover, and install modular packages from the open agent skills ecosystem using the Skills CLI.
video-edit
338.7KUsers of the RunComfy CLI who need to automate video editing tasks like restyling, background swapping, or motion transfer
The video-edit skill acts as a router for the RunComfy CLI, selecting between Wan 2.7 Edit-Video, Kling 2.6 Pro Motion Control, and Lucy Edit Restyle models based on user intent to perform video transformations.
lark-doc
305.2KUsers who need to automate document management, content updates, and media handling within the Lark/Feishu ecosystem
The lark-doc skill enables agents to read, create, and edit Lark (Feishu) documents, including Docx and Wiki formats. It supports content manipulation via XML or Markdown, media handling, and resource management for document covers. The skill integrates with other Lark skills by identifying and delegating operations for embedded objects like spreadsheets, databases, and mind notes.
ai-image-generation
229.9KDevelopers and AI agents requiring programmatic access to diverse image generation and editing models through a unified CLI
The AI Image Generation skill provides a command-line interface for generating and editing images using the RunComfy platform. It routes requests across 11+ AI models, including FLUX 2, Google Nano Banana, OpenAI GPT Image 2, ByteDance Seedream, Alibaba Qwen, and Wan 2-7. The skill selects the appropriate model based on user intent, such as typography precision, photorealism, or speed, and executes the generation via the runcomfy CLI.