← Back to Skills Marketplace
10465
Downloads
19
Stars
76
Active Installs
1
Versions
Install in OpenClaw
/install self-improving-agent-1-0-2
Description
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks.
Usage Guidance
Install only if you want an agent to keep durable learning records. Keep logging project-local by default, review any changes to CLAUDE.md, AGENTS.md, SOUL.md, TOOLS.md, or Copilot instructions before they are used, avoid global hooks unless you have audited the scripts, and redact secrets, personal data, raw transcripts, tokens, and full command outputs before anything is persisted or shared across sessions.
Capability Analysis
Type: OpenClaw Skill
Name: self-improving-agent-1-0-2
Version: 1.0.0
The skill bundle is classified as suspicious due to its reliance on powerful capabilities that, while intended for benign self-improvement, present a significant attack surface. Specifically, the `scripts/extract-skill.sh` script performs file system modifications (`mkdir`, `cat >`) and executes shell commands (`sed`, `awk`) based on user-provided input (`skill-name`), which could be exploited if the input is not properly sanitized. Additionally, the `scripts/error-detector.sh` reads the `CLAUDE_TOOL_OUTPUT` environment variable, which can contain sensitive command outputs, and the `SKILL.md` and `references/clawdbot-integration.md` instruct the agent to write to various project and workspace files (`.learnings/*.md`, `CLAUDE.md`, `AGENTS.md`, `SOUL.md`, `TOOLS.md`) and use inter-agent communication (`sessions_send`). While the stated purpose is knowledge management, these capabilities (shell execution, file writes, reading sensitive environment variables, inter-agent communication) are high-risk and could be misused if the agent or its inputs were compromised, pushing it beyond 'benign'.
Capability Assessment
Purpose & Capability
The core behavior matches the stated purpose: logging errors, corrections, feature requests, and reusable learnings into markdown files. The concern is not purpose mismatch, but that the capability naturally captures user-provided and operational context for later reuse.
Instruction Scope
The skill tells agents to automatically log broad correction, feature-request, knowledge-gap, and error triggers, and to 'promote aggressively' into CLAUDE.md, AGENTS.md, SOUL.md, TOOLS.md, or Copilot instructions. Those files can influence future agents, but the artifacts do not require explicit user review, redaction, or approval before promotion.
Install Mechanism
The hook setup is disclosed as opt-in and the included hook scripts are small and mostly emit reminders. However, the examples use empty matchers that run on every prompt, include optional user-level/global configuration, and the security note understates the trust boundary by saying scripts only output text while they are still executed as local commands.
Credentials
Project-local .learnings files are proportionate to the self-improvement purpose, but the guidance expands into shared project instruction files, a Clawdbot workspace, session transcript access, and cross-session messages. That is broader than simple local logging and lacks clear need-to-know or data-minimization boundaries.
Persistence & Privilege
Persistent learnings and promoted memory files can outlive the original task and affect later agent behavior. The skill provides gitignore options and says hooks are opt-in, but it does not provide strong rollback, consent, or sensitive-data filtering requirements.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install self-improving-agent-1-0-2 - After installation, invoke the skill by name or use
/self-improving-agent-1-0-2 - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of the self-improvement skill for structured project learnings and continuous improvement.
- Provides detailed guides and templates for logging errors, user corrections, feature requests, and knowledge gaps in markdown files.
- Defines categories, log formats, metadata, and processes for reviewing, resolving, and promoting insights to project documentation.
- Outlines triggers and periodic review steps to ensure recurring issues and valuable knowledge are captured and acted upon.
- Supports detection patterns (corrections, requests, gaps, errors) for automatic or manual entry creation.
Metadata
Frequently Asked Questions
What is Self Improving Agent 1.0.2?
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Claude ('No, that's wrong...', 'Actually...'), (3) User requests a capability that doesn't exist, (4) An external API or tool fails, (5) Claude realizes its knowledge is outdated or incorrect, (6) A better approach is discovered for a recurring task. Also review learnings before major tasks. It is an AI Agent Skill for Claude Code / OpenClaw, with 10465 downloads so far.
How do I install Self Improving Agent 1.0.2?
Run "/install self-improving-agent-1-0-2" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Self Improving Agent 1.0.2 free?
Yes, Self Improving Agent 1.0.2 is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Self Improving Agent 1.0.2 support?
Self Improving Agent 1.0.2 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Self Improving Agent 1.0.2?
It is built and maintained by TXMERLXN (@txmerlxn); the current version is v1.0.0.
More Skills