18816132863 Self-Improvement Skill
Log learnings and errors to markdown files for continuous improvement. Coding agents can later process these into fixes, and important learnings get promoted to project memory.
⚠️ CRITICAL: Session Start Routine
Every session MUST start by reading learning files to avoid repeating mistakes.
OpenClaw Setup (Automatic)
Add this to your AGENTS.md to ensure learnings are loaded every session:
## Every Session
Before doing anything else:
1. Read `SOUL.md` — this is who you are
2. Read `USER.md` — this is who you're helping
3. Read `memory/YYYY-MM-DD.md` (today + yesterday) for recent context
4. **If in MAIN SESSION**: Also read `MEMORY.md`
5. **Read `.learnings/LEARNINGS.md` and `.learnings/ERRORS.md`** — learn from past mistakes!
Don't ask permission. Just do it.
Why This Matters
| Without Session Loading | With Session Loading |
|---|---|
| ❌ Repeats same mistakes | ✅ Avoids known errors |
| ❌ Forgets optimizations | ✅ Applies learned improvements |
| ❌ No continuity | ✅ Continuous improvement |
Quick Reference
| Situation | Action |
|---|---|
| Command/operation fails | Log to .learnings/ERRORS.md |
| User corrects you | Log to .learnings/LEARNINGS.md with category correction |
| User wants missing feature | Log to .learnings/FEATURE_REQUESTS.md |
| API/external tool fails | Log to .learnings/ERRORS.md with integration details |
| Knowledge was outdated | Log to .learnings/LEARNINGS.md with category knowledge_gap |
| Found better approach | Log to .learnings/LEARNINGS.md with category best_practice |
| Broadly applicable learning | Promote to CLAUDE.md, AGENTS.md, or TOOLS.md |
Workspace Structure
~/.openclaw/workspace/
├── AGENTS.md # ⭐ Add session start routine here!
├── SOUL.md # Behavioral guidelines
├── USER.md # User preferences
├── TOOLS.md # Tool configurations
├── MEMORY.md # Long-term memory
├── memory/ # Daily memory files
│ └── YYYY-MM-DD.md
└── .learnings/ # This skill's log files
├── LEARNINGS.md # Best practices, corrections
├── ERRORS.md # Failures and fixes
└── FEATURE_REQUESTS.md
Installation
Via SkillHub (recommended):
skillhub install self-improving-agent
Manual:
git clone https://github.com/peterskoett/self-improving-agent.git ~/.openclaw/workspace/skills/self-improving-agent
Post-Installation Setup
- Create learning files:
mkdir -p ~/.openclaw/workspace/.learnings
touch ~/.openclaw/workspace/.learnings/LEARNINGS.md
touch ~/.openclaw/workspace/.learnings/ERRORS.md
touch ~/.openclaw/workspace/.learnings/FEATURE_REQUESTS.md
- Add session start routine to AGENTS.md (see above)
Logging Format
Learning Entry
Append to .learnings/LEARNINGS.md:
## [LRN-YYYYMMDD-XXX] category
**Logged**: ISO-8601 timestamp
**Priority**: low | medium | high | critical
**Status**: pending | resolved | promoted
**Area**: frontend | backend | infra | docs | config
### Summary
One-line description of what was learned
### Details
Full context: what happened, what was wrong, what's correct
### Suggested Action
Specific fix or improvement to make
### Metadata
- Source: conversation | error | user_feedback
- Related Files: path/to/file.ext
- Tags: tag1, tag2
---
Error Entry
Append to .learnings/ERRORS.md:
## [ERR-YYYYMMDD-XXX] skill_or_command_name
**Logged**: ISO-8601 timestamp
**Priority**: high
**Status**: pending | resolved
**Area**: frontend | backend | infra | docs | config
### Summary
Brief description of what failed
### Error
Actual error message or output
### Context
- Command/operation attempted
- Environment details if relevant
### Suggested Fix
What might resolve this
### Resolution (when fixed)
- **Resolved**: timestamp
- **Notes**: What was done
---
Promotion Targets
When learnings prove broadly applicable, promote them:
| Learning Type | Promote To | Example |
|---|---|---|
| Behavioral patterns | SOUL.md |
"Be concise, avoid disclaimers" |
| Workflow improvements | AGENTS.md |
"Spawn sub-agents for long tasks" |
| Tool gotchas | TOOLS.md |
"PDF font path: /usr/share/fonts/..." |
| User preferences | USER.md |
"Show skill usage after each reply" |
How to Promote
- Distill the learning into a concise rule
- Add to appropriate file
- Update original entry:
**Status**: promoted
Detection Triggers
Automatically log when you notice:
| Trigger | Action |
|---|---|
| "No, that's wrong..." | Log as correction |
| Command fails | Log to ERRORS.md |
| Found better way | Log as best_practice |
| User says "remember this" | Log and promote |
Best Practices
- Log immediately - context is freshest right after the issue
- Be specific - future agents need to understand quickly
- Promote aggressively - if broadly useful, add to workspace files
- Read at session start - avoid repeating mistakes
- Link related entries - use
See Alsofor recurring issues
Example: Complete Workflow
1. Error Occurs
PDF generated with Chinese text shows as garbled characters
2. Log to ERRORS.md
## [ERR-20260317-001] pdf_font_error
**Logged**: 2026-03-17T09:40:00Z
**Priority**: high
**Status**: pending
**Area**: docs
### Summary
PDF Chinese text shows as garbled characters due to wrong font path
### Error
Font path /usr/share/fonts/truetype/wqy/wqy-zenhei.ttc does not exist
### Suggested Fix
Use correct font path: /usr/share/fonts/wqy-microhei/wqy-microhei.ttc
---
3. Fix Applied, Update Entry
### Resolution
- **Resolved**: 2026-03-17T09:43:00Z
- **Notes**: Corrected font path, PDF now displays Chinese correctly
4. Promote to TOOLS.md
### PDF Generation
- **Chinese font path**: `/usr/share/fonts/wqy-microhei/wqy-microhei.ttc`
5. Update Learning Status
**Status**: promoted
Summary
| Step | Action |
|---|---|
| 1️⃣ | Session Start: Read .learnings/ files |
| 2️⃣ | On Error: Log to ERRORS.md |
| 3️⃣ | On Learning: Log to LEARNINGS.md |
| 4️⃣ | When Fixed: Update status to resolved |
| 5️⃣ | When Broad: Promote to workspace files |
For full documentation, see the original SKILL.md or visit: https://github.com/peterskoett/self-improving-agent