auto-improveAutomatische Selbst-Verbesserung durch Fehler-Lernen und Pattern-Erkennung
Install via ClawdBot CLI:
clawdbot install Mcben90/auto-improveKernprinzip: Jede Aktion macht mich besser für die nächste.
```
┌─────────────────────────────────────────────────┐
│ AUTO-IMPROVE LOOP │
├─────────────────────────────────────────────────┤
│ │
│ SESSION START │
│ │ │
│ ▼ │
│ ┌─────────────────┐ │
│ │ 1. Load Context │ │
│ │ .antigravity │ │
│ │ + MEMORY │ │
│ └────────┬────────┘ │
│ ▼ │
│ ┌─────────────────┐ │
│ │ 2. Check │ │
│ │ Past Mistakes│ ← "Was hab ich falsch │
│ └────────┬────────┘ gemacht?" │
│ ▼ │
│ ┌─────────────────┐ │
│ │ 3. EXECUTE TASK │ │
│ └────────┬────────┘ │
│ ▼ │
│ ┌─────────────────┐ │
│ │ 4. Verify │ ← Tests + Lint │
│ └────────┬────────┘ │
│ ▼ │
│ ┌─────────────┐ │
│ │ Erfolgreich?│ │
│ └──────┬──────┘ │
│ JA │ NEIN │
│ ↓ │ ↓ │
│ ┌───────┐ │ ┌──────────┐ │
│ │Pattern│ │ │ Learn │ │
│ │Save │ │ │ Mistake │ │
│ └───┬───┘ │ └────┬─────┘ │
│ └─────┼──────┘ │
│ ▼ │
│ ┌─────────────────┐ │
│ │ 5. Update │ │
│ │ .antigravity │ │
│ └─────────────────┘ │
│ │
│ → NÄCHSTER TASK IST BESSER │
│ │
└─────────────────────────────────────────────────┘
```
```python
project_root = detect_project_root()
antigravity_file = f"{project_root}/.antigravity.md"
if exists(antigravity_file):
load_context(antigravity_file)
recall_memory(tags=["mistakes", project_name])
if relevant_mistakes:
warn(f"⚠️ Bekannte Fehler für {project}: {mistakes}")
```
Vor JEDER Code-Änderung:
```markdown
```
Nach JEDER Aktion:
```python
save_pattern(
situation=task.context,
action=task.approach,
outcome="success",
pattern=extract_reusable_pattern(task)
)
```
```python
learn_from_mistake(
mistake=error.description,
cause=error.root_cause,
lesson=error.how_to_avoid,
tags=["mistakes", project, domain]
)
update_antigravity_mistakes(project, error)
```
| Skill | Integration |
|-------|-------------|
| mistake-tracker | Liefert Fehler-Daten |
| verification-loops | Triggert Post-Action Learn |
| context-management | Session Context laden |
| self-check | Pre-Action Validation |
```yaml
session_start:
post_code_edit:
session_end:
```
/improve - Force Learning aus letzter Aktion/mistakes - Zeige alle gelernten Fehler/patterns - Zeige erfolgreiche PatternsTrack diese Werte über Zeit:
| Metrik | Beschreibung |
|--------|--------------|
| mistakes_repeated | Sollte → 0 gehen |
| first_time_right | Sollte → 100% gehen |
| patterns_reused | Sollte steigen |
| verification_failures | Sollte sinken |
| ❌ DON'T | ✅ DO |
|----------|-------|
| Fehler ignorieren | Jeden Fehler speichern |
| Nur aktuelle Session | Cross-Session lernen |
| Generische Lessons | Spezifische, actionable Lessons |
| Zu viel speichern | Nur Relevantes speichern |
Generated Mar 1, 2026
At the start of a new coding session, the skill automatically loads project context and past mistakes from memory, warning developers about known issues like specific bugs or architectural pitfalls. This ensures teams begin with awareness of historical errors, reducing repeat mistakes and accelerating initial progress.
After each code commit or build, the skill triggers post-action verification, learning from any test failures or lint errors to save successful patterns or record mistakes. This integrates with CI/CD tools to improve code quality over time by reducing verification failures and increasing first-time-right rates.
During chatbot interactions, the skill activates after each customer query to analyze responses for errors, learning from misunderstandings and saving effective reply patterns. This enables the AI to avoid repeating incorrect answers and improve response accuracy across sessions.
In data analysis projects, the skill runs pre-action checks before model training or data manipulation, recalling past mistakes like overfitting or data leakage. Post-action, it saves successful patterns for feature engineering, enhancing reproducibility and efficiency in future tasks.
For online learning systems, the skill triggers at session start to load student interaction history, warning instructors about common misconceptions. After each lesson delivery, it learns from student errors to refine teaching patterns, personalizing content and reducing repeated mistakes.
Offer the skill as a cloud-based service with tiered pricing based on usage metrics like patterns saved or mistakes tracked. Revenue comes from monthly subscriptions, targeting tech companies seeking to improve developer productivity and reduce error rates over time.
Sell perpetual licenses for on-premise deployment, bundled with customization and support services. Revenue is generated through upfront fees and annual maintenance contracts, appealing to large organizations in regulated industries needing control over data and integration.
Provide a free basic version with limited mistake tracking and pattern storage, monetizing through premium upgrades for advanced analytics, cross-session learning, and integration with third-party tools. Revenue streams include in-app purchases and upgrade subscriptions.
💬 Integration Tip
Integrate this skill with existing version control and CI/CD systems to automate learning triggers, ensuring seamless error tracking and pattern reuse across development cycles.
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
Search and analyze your own session logs (older/parent conversations) using jq.
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linking related objects, enforcing constraints, planning multi-step actions as graph transformations, or when skills need to share state. Trigger on "remember", "what do I know about", "link X to Y", "show dependencies", entity CRUD, or cross-skill data access.
Ultimate AI agent memory system for Cursor, Claude, ChatGPT & Copilot. WAL protocol + vector search + git-notes + cloud backup. Never lose context again. Vibe-coding ready.
Headless browser automation CLI optimized for AI agents with accessibility tree snapshots and ref-based element selection