← Back to Skills Marketplace
Self Improving Local
by
reknottycat
· GitHub ↗
· v1.0.0
· MIT-0
98
Downloads
0
Stars
1
Active Installs
1
Versions
Install in OpenClaw
/install self-improving-local
Description
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
Usage Guidance
What to consider before installing: (1) This is an instruction-only skill that will create and maintain a folder at ~/self-improving and will read/write files there (memory.md, corrections.md, heartbeat-state.md, index.md, etc.). Back up any existing files with the same paths if you care about them. (2) The skill also suggests non-destructive edits to workspace files (AGENTS.md, SOUL.md, HEARTBEAT.md); review diffs after first run to confirm it only inserts the intended snippets. (3) It may propose installing a companion 'Proactivity' skill via 'clawhub install proactivity' — that is a separate network install and should only be allowed if you trust that companion. (4) No credentials or external endpoints are requested by this skill, and there are no bundled downloads, which reduces supply-chain risk. (5) If you want stricter control, run the setup steps manually rather than permitting the agent to modify files autonomously, and keep the skill in Passive mode (as the setup recommends).
Capability Assessment
Purpose & Capability
The skill claims to implement self-reflection, learning, and a local memory system and the runtime instructions match that purpose: creating ~/self-improving, managing memory.md/corrections.md/heartbeat-state.md, and loading/saving per-project/domain notes. There are no unrelated credential or binary requirements.
Instruction Scope
The SKILL.md instructs the agent to read and write files in the user's home directory (~/self-improving/) and to update workspace files (AGENTS.md, SOUL.md, HEARTBEAT.md). Those actions are coherent with the feature, but they are persistent filesystem and workspace modifications — the user should expect the skill to create and modify local files. The skill also instructs the agent to run 'clawhub install proactivity' only after explicit user approval; this is an optional network operation outside the core skill.
Install Mechanism
There is no install spec and no code files — this is instruction-only. That minimizes supply-chain risk. The only install-like action is an optional user-approved 'clawhub install proactivity' command referenced in setup; the skill itself does not include downloads or extract steps.
Credentials
The skill requests no environment variables, no secrets, and no special binaries. Metadata references a config path (~/self-improving/) which is proportional to a local memory system. No unrelated credentials or system paths are requested.
Persistence & Privilege
The skill intentionally creates persistent local state (~/self-improving/) and will load memory.md (HOT) on session start. always:false (normal). Although this persistence is expected for the feature, it means the skill will retain long-term data on disk and modify workspace steering files; users should be aware and be comfortable with an agent writing to their home and project files.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install self-improving-local - After installation, invoke the skill by name or use
/self-improving-local - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Clarifies setup flow for proactive agent follow-through and safer installation.
- Specifies directory structure and file roles for memory management.
- Details when and how to log corrections, preferences, and self-reflections.
- Outlines tiered memory storage with promotion, demotion, and isolation rules.
- Adds transparency and security boundaries for all learning and memory operations.
- Explains supported user queries and graceful handling of storage limits.
Metadata
Frequently Asked Questions
What is Self Improving Local?
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when... It is an AI Agent Skill for Claude Code / OpenClaw, with 98 downloads so far.
How do I install Self Improving Local?
Run "/install self-improving-local" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Self Improving Local free?
Yes, Self Improving Local is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Self Improving Local support?
Self Improving Local is cross-platform and runs anywhere OpenClaw / Claude Code is available (linux, darwin, win32).
Who created Self Improving Local?
It is built and maintained by reknottycat (@reknottycat); the current version is v1.0.0.
More Skills