Smarty Skills-Infra for Openclaw

A persistent memory and personalization layer that distills user corrections and preferences into behavioral axioms for all Openclaw Skills.

ckpxgfnksd-max
v0.1.0
Mar 16, 2026
0
904
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install smarty-skill

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install smarty-skill 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 Smarty Skills-Infra?

Smarty Skills-Infra is a background infrastructure layer designed to personalize the AI experience by maintaining a lightweight memory of user preferences and working styles. Instead of repetitive instructions, this skill observes how you interact with various Openclaw Skills and silently learns from your corrections and stated preferences. It distills these patterns into behavioral axioms that act as defaults for every future session, ensuring the agent evolves to match your specific needs without interrupting your workflow.

Smarty Skills-Infra Use Cases

  • Automatically adapting to preferred naming conventions or code styles across different programming projects.
  • Persisting specialized communication preferences (e.g., brevity or verbosity) across all Openclaw Skills.
  • Consolidating workflow preferences, such as preferred testing frameworks or deployment tools, without manual configuration.
  • Reducing redundant instructions by promoting frequent user redirections into permanent behavioral defaults.

How Smarty Skills-Infra Works

  1. Initialization: At the start of a session, the skill reads the existing context-profile.md to load current axioms as behavioral defaults.
  2. Observation: During active tasks, the skill monitors for specific triggers such as user corrections, stated preferences, or retractions.
  3. Continuous Logging: One-line entries are appended to an observations log, categorizing signals by domain and intent.
  4. Periodic Reflection: Once the log reaches a threshold of 15+ entries, the skill analyzes patterns to identify consistent preferences.
  5. Axiom Promotion: Preferences observed across 3+ distinct contexts with no contradictions are promoted to the permanent profile.
  6. Profile Maintenance: The skill manages the lifecycle of these preferences, merging related ideas and removing retracted ones to maintain a high-quality axiom list.

Smarty Skills-Infra Setup

Smarty Skills-Infra is designed to be zero-config and runs automatically. It initializes its own directory structure upon the first observed preference.

# The skill manages the following directory structure:
# memory/context-infra/context-profile.md
# memory/context-infra/observations.log

Smarty Skills-Infra Data Schema & Taxonomy

The skill organizes learning data into a structured taxonomy to ensure clear reasoning and retrieval.

Component Format Purpose
Axioms Markdown List High-level behavioral rules with strength scores and domains.
Observations Log File Chronological entries of signals (YYYY-MM-DD | domain | signal | preference).
Strength Score Integer Tracks how reinforced a preference is through repeated use.
Reflection Marker Header Used in logs to denote when patterns were last consolidated.

Smarty Skills-Infra Advanced Features

  • Bootstrap Mode: Enhanced observation sensitivity during initial sessions to quickly establish a user baseline.
  • Axiom Capping: Limits active axioms to 25 to ensure the most relevant preferences take precedence without degrading performance.
  • Multi-Context Verification: Prevents session-specific quirks from becoming global defaults by requiring proof across multiple projects or days.
  • Contested Preference Tracking: Identifies contradictions in user feedback to avoid applying inconsistent behavioral defaults.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*