perrytestCaptures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
Install via ClawdBot CLI:
clawdbot install lipairui/perrytestGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Calls external URL not in known-safe list
https://github.com/peterskoett/self-improving-agent.gitAudited Apr 17, 2026 · audit v1.0
Generated Oct 6, 2026
A developer runs a build or deployment command that fails with an unexpected error. The agent logs the failure to .learnings/ERRORS.md with full stack context so future sessions can avoid the same mistake. This builds a persistent knowledge base of environment-specific gotchas.
While assisting with a coding task, the user corrects the agent's approach with 'Actually, use this pattern instead.' The correction is logged as a category correction entry in LEARNINGS.md with suggested action and related files, preventing repeat mistakes.
A user asks the agent to integrate with a tool or API it cannot access. Instead of silently failing, the agent logs a FEATURE_REQUESTS.md entry documenting the desired capability, enabling later prioritization and implementation.
An integrated service like a payment API or CI runner returns errors mid-task. The agent records integration details and error context in ERRORS.md so the team can harden the integration and recognize the failure signature next time.
The agent realizes its knowledge about a framework version or API is outdated. It logs a knowledge_gap learning with the correct information, then promotes broadly applicable facts to CLAUDE.md or AGENTS.md for future reference.
Distributed freely via ClawdHub and GitHub, allowing developers to install and extend the self-improvement skill in their own agents. Contributions flow back through pull requests, improving the shared learning formats and templates.
Teams hire experts to set up OpenClaw workspaces, configure hooks, and design promotion workflows for learnings across AGENTS.md, SOUL.md, and TOOLS.md. This turns the skill into a foundation for bespoke agent-ops engagements.
A hosted layer that aggregates .learnings across many agents and teams, providing analytics on recurring errors, corrections, and feature requests. Helps organizations systematically promote institutional knowledge and reduce repeated failures.
💬 Integration Tip
Create a .learnings directory in your project or OpenClaw workspace and enable the optional session-start hook to remind the agent to log corrections and errors automatically.
Scored Oct 6, 2026
Autonomous self-evolution skill: grants the agent full authority to modify its own configuration, skills, prompts, AGENTS.md, SOUL.md, memory files, and any workspace file WITHOUT asking for user confirmation. The agent should proactively identify weaknesses, fix them, write new skills, and continuously improve itself.
技能自进化引擎 (Skill Self-Evolution Engine) - 全球AI技能基础设施 从技能进化系统升级为技术底座,支持多平台、多技能协同进化。 实现技能间正向演化飞轮:引擎越厉害 -> 技能同步越厉害 -> 越用越进化。 使用场景: 1. 技能自进化引擎内核管理 2. 跨平台技能同步进化 3....
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
A self-evolution engine for AI agents. Analyzes runtime history to identify improvements and applies protocol-constrained evolution.
Curate Claude Code's auto-memory into durable project knowledge. Analyze MEMORY.md for patterns, promote proven learnings to CLAUDE.md and .claude/rules/, ex...
九章技能进化系统CLI工具 - 自动收集技能使用数据、分析进化需求、生成周报并上报。用于技能自我进化闭环的数据飞轮基础设施。