self▌

$22

Captures learnings, errors, and corrections to enable continuous improvement across sessions.

Skill by ara.so — Daily 2026 Skills collection.

Universal self-improving agent that learns from all skill experiences using multi-memory architecture. \n \n Implements semantic, episodic, and working memory to extract patterns, abstract insights, and continuously evolve skill guidance across the codebase \n Auto-triggers on skill completion, errors, and session events via hooks-based integration; detects and corrects inaccurate guidance with traceable evolution markers \n Prioritizes updates across 10+ skill categories (PRD planning, architec

Autonomous design critique mode that adds annotations to web pages via the Agentation toolbar in a visible browser. \n \n Requires the Agentation toolbar installed on the target page and the agent-browser skill available; launches in headed mode so users watch the agent scan, hover, and annotate in real time \n Uses coordinate-based mouse events to trigger annotation dialogs (standard element clicks do not work); includes helper patterns for scrolling, bounding box lookup, and dialog interaction

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.

元认知自学习系统 - 基于预测编码和价值驱动的Skill自动演化。

When to use this skill:

Automated learning capture in CI pipelines that deduplicates failure patterns and proposes prevention rules. \n \n Inspects PR check results and CI failures to identify recurring patterns tracked by stable pattern_key , promoting only when recurrence thresholds are met (3+ occurrences across 2+ distinct runs within 30 days) \n Ingests learning candidates from simplify-and-harden-ci and emits machine-readable YAML output without interactive prompts, suitable for headless GitHub Actions workflows

4-tier autonomous self-healing system for OpenClaw Gateway with Claude Code as AI emergency doctor. \n \n Escalates through watchdog monitoring, HTTP health checks with retries, AI-powered diagnosis via Claude Code, and Discord/Telegram alerts for human intervention \n Captures persistent learning documentation (symptom to solution mappings) and reasoning logs for explainable AI decision-making \n Includes metrics dashboard for tracking recovery success rates, timing, and trends across incidents