← Back to Skills Marketplace
90
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install self-evolving-skills
Description
Agent 自我进化机制 - 将成功的复杂任务沉淀为可复用的 Skill，实现持续学习和知识积累。
Usage Guidance
This skill appears to do what it says: it helps an agent save workflows as SKILL.md and optionally scripts/templates. Before installing or enabling it, consider: (1) verify that your agent runtime enforces the 'ask user' confirmation before any file writes, (2) restrict the skill storage directory to a non-privileged location you control (e.g., a dedicated ./skills or ~/.agent/skills path), (3) inspect any scripts the agent generates before running them, and (4) consider limiting autonomous agent actions or requiring interactive approval for creating executable files. If you want stronger assurance, run static checks on created scripts (linters, virus scan) and avoid granting blanket permission for the agent to create/execute code.
Capability Analysis
Type: OpenClaw Skill
Name: self-evolving-skills
Version: 1.0.1
The skill bundle defines a 'self-evolution' mechanism for an AI agent to document and save successful workflows as reusable skills. It includes a Python validation script (scripts/skill_validator.py), templates, and comprehensive documentation on decision-making and formatting. While the skill instructs the agent to perform file system operations (creating directories and writing files), the instructions emphasize user confirmation and organizational structure without any evidence of malicious intent, data exfiltration, or unauthorized persistence.
Capability Assessment
Purpose & Capability
Name/description describe an agent mechanism for converting successful tasks into reusable skills; the files and included validator script are appropriate for that purpose. There are no unexpected env vars, binaries, or external service credentials requested.
Instruction Scope
SKILL.md explicitly instructs the agent to create files and optional scripts/templates/references in user-specified skill directories and to ask the user for confirmation before creating a Skill. This is within scope, but the instructions allow the agent to produce executable code (scripts/) — which is expected for a 'save script' flow but expands the attack surface if created files are later executed without review.
Install Mechanism
No install spec and no network/downloads — instruction-only plus included helper script. No high-risk install behavior (no remote archive downloads or package installs).
Credentials
The skill requires no environment variables, credentials, or config paths. The included Python validator reads local skill files (expected for validation) but does not request secrets or unrelated credentials.
Persistence & Privilege
always:false (normal). The skill enables creation of persistent files under user-chosen directories; combined with model invocation allowed by default, that means an agent could autonomously propose creating code artifacts — mitigated by explicit 'ask user confirmation' steps in the instructions, but users should ensure prompts are actually enforced by their agent runtime.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install self-evolving-skills - After installation, invoke the skill by name or use
/self-evolving-skills - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
v1.0.1
- No file changes detected in this release.
- Version and documentation remain unchanged from v1.0.0.
- No user-facing updates in this version.
v1.0.0
- Initial release of **self-evolving-skills** (version 1.0.0).
- Defines a mechanism for agents to turn successful complex tasks into reusable, procedural Skills for ongoing learning and knowledge accumulation.
- Outlines clear criteria for when new Skills should or should not be created, and provides a step-by-step decision and confirmation process.
- Provides detailed guidelines for Skill formatting, directory structure, content standards, and file operation instructions—tool-agnostic for broad compatibility.
- Includes maintenance and lifecycle best practices to ensure Skills remain accurate and valuable over time.
Metadata
Frequently Asked Questions
What is self-evolving-skills?
Agent 自我进化机制 - 将成功的复杂任务沉淀为可复用的 Skill，实现持续学习和知识积累。 It is an AI Agent Skill for Claude Code / OpenClaw, with 90 downloads so far.
How do I install self-evolving-skills?
Run "/install self-evolving-skills" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is self-evolving-skills free?
Yes, self-evolving-skills is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does self-evolving-skills support?
self-evolving-skills is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created self-evolving-skills?
It is built and maintained by tunsuy (@tunsuy); the current version is v1.0.1.
More Skills