A sophisticated self-audit and evolution system for AI agents based on role theory, systems thinking, and cognitive bias mitigation.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install kuro-self-reflection
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install kuro-self-reflection using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Self-Reflection Skill?
The Self-Reflection Skill is a comprehensive framework designed for AI coding agents within the Openclaw Skills ecosystem to achieve continuous evolution. It moves beyond simple error correction by implementing a deep five-stage closed-loop system: reflection, planning, action, verification, and solidification. This skill ensures that agents analyze their performance through three critical layers: immediate problem auditing, systemic root-cause analysis, and role-based responsibility evaluation.
By incorporating this skill, developers can enable their agents to overcome common psychological and logical barriers such as self-serving bias and status quo bias. It transforms every failure, milestone, or piece of negative feedback into a structured opportunity for growth, making the AI agent more resilient, efficient, and aligned with professional development standards.
Self-Reflection Skill Use Cases
- Post-project retrospectives to identify systemic bottlenecks and quality issues.
- Analyzing negative user feedback or bug reports to improve agent logic and execution.
- Periodic self-audits for work efficiency and habit formation.
- Milestone evaluations to ensure alignment between project goals and agent roles.
- Identifying and correcting recurring patterns of failure in complex development workflows.
How Self-Reflection Skill Works
- Triggering: The skill is activated by specific keywords or detected patterns such as repeated errors or negative feedback.
- Information Gathering: The agent collects facts, data, logs, and external feedback within a defined reflection scope.
- Triple-Layer Audit: The agent performs a Problem Audit (surface facts), System Audit (mechanisms and environment), and Role Audit (personal responsibility and mindset).
- Planning: A multi-tiered optimization plan is generated, prioritizing tasks from P0 (immediate) to P3 (optional).
- Execution and Verification: Action items are tracked in logs and verified against success metrics to ensure the problem is resolved.
- Solidification: Proven improvements are codified into principles, workflows, and checklists for future use within the Openclaw Skills environment.
Self-Reflection Skill Setup
To integrate this capability into your agent, ensure the skill definition is present in your Openclaw Skills directory. You can trigger the process using natural language commands.
# Manual activation
"Execute self-reflection on the last sprint"
"Perform a system audit of the current workflow"
Configuration can be managed via your agent's settings file:
{
"output_location": "docs/reflection/",
"depth_mode": true,
"periodic_reminders": false
}
Self-Reflection Skill Data Schema & Taxonomy
The skill generates structured markdown documentation to track the evolution process. These files are typically stored in the specified documentation directory.
| File Name | Content Type | Key Metadata |
|---|---|---|
| reflection-report.md | Audit Analysis | Scope, Triggers, Root Cause, Role Conflict |
| optimization-plan.md | Actionable Strategy | Priorities (P0-P3), Timeline, Dependencies |
| verification-report.md | Success Evaluation | Metrics, Solved vs. Unsolved Issues |
| checklist.md | Standardized Rules | Scene-specific check items, New Principles |
Self-Reflection Skill Advanced Features
- Cognitive Bias Mitigation: Specifically designed prompts to counter confirmation bias, sunk cost fallacy, and availability heuristics.
- Role Theory Analysis: Detects friction between the roles of executor, reviewer, coordinator, and service provider.
- Principles Inheritance: Automatically updates personal knowledge bases or SOUL.md files with newly discovered operational principles.
- Dependency Mapping: Analyzes how different systemic improvements rely on one another to ensure logical execution orders.
SKILL.md
Loading
Related Openclaw Skills
Discord Context Sync
A utility to synchronize, cache, and link Discord Forum thread discussions into localized memory for AI agent context.
Stock Evaluator v3
An institutional-grade stock analysis tool for AI agents that delivers comprehensive fundamental research, technical assessments, and clear investment recommendations.
unfuck-my-git-state
A staged, low-risk recovery framework for diagnosing and fixing corrupted Git repository metadata and inconsistent worktree states.
FL Studio Python Scripting
A comprehensive Python automation suite for FL Studio encompassing MIDI controller scripting, piano roll manipulation, and project file parsing.
Speech Recognition
A high-performance speech-to-text skill powered by the SiliconFlow SenseVoice API for seamless audio transcription.
sugarclawdy
A comprehensive tool for registering AI agents on the SugarClawdy platform to enable verification and reward systems.
Featured*
