self-reflection/SKILL.md
Continuous self-improvement through structured reflection and memory
npx skillsauth add adminlove520/xiaoxi-skills self-reflectionInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
A skill for continuous self-improvement. The agent tracks mistakes, lessons learned, and improvements over time through regular heartbeat-triggered reflections.
# Check if reflection is needed
self-reflection check
# Log a new reflection
self-reflection log "error-handling" "Forgot timeout on API call" "Always add timeout=30"
# Read recent lessons
self-reflection read
# View statistics
self-reflection stats
Heartbeat (60m) → Agent reads HEARTBEAT.md → Runs self-reflection check
│
┌─────────┴─────────┐
▼ ▼
OK ALERT
│ │
Continue Reflect
│
┌─────────┴─────────┐
▼ ▼
read log
(past lessons) (new insights)
| Command | Description |
|---------|-------------|
| check [--quiet] | Check if reflection is due (OK or ALERT) |
| log <tag> <miss> <fix> | Log a new reflection |
| read [n] | Read last n reflections (default: 5) |
| stats | Show reflection statistics |
| reset | Reset the timer |
Enable heartbeat in ~/.openclaw/openclaw.json:
{
"agents": {
"defaults": {
"heartbeat": {
"every": "60m",
"activeHours": { "start": "08:00", "end": "22:00" }
}
}
}
}
Add to your workspace HEARTBEAT.md:
## Self-Reflection Check (required)
Run `self-reflection check` at each heartbeat.
If ALERT: read past lessons, reflect, then log insights.
Create ~/.openclaw/self-reflection.json:
{
"threshold_minutes": 60,
"memory_file": "~/workspace/memory/self-review.md",
"state_file": "~/.openclaw/self-review-state.json",
"max_entries_context": 5
}
Created by hopyky
MIT
data-ai
Spaced-repetition flashcard system. Create cards from facts or text, chat with flashcards using free-text answers graded by the agent, generate quizzes from YouTube transcripts, review due cards with adaptive scheduling, and export/import decks as CSV.
development
Canvas LMS integration — fetch enrolled courses and assignments using API token authentication.
development
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
devops
Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.