save-learning

Converts user instructions into persistent typed learning stored via Caliber.

1|Updated Aug 22, 2025
One-click install
npx skills add https://github.com/interserver/teams-chat-bot --skill save-learning-interserver
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: save-learning
Source: https://github.com/interserver/teams-chat-bot/tree/main/.cursor/skills/save-learning
Command: npx skills add https://github.com/interserver/teams-chat-bot --skill save-learning-interserver

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents users from having to restate the same instructions or preferences every time by converting recurring requests into persistent project learnings.

Core Features & Use Cases

  • Captures repeatable instructions: Detects phrases like “remember this”, “always do X”, and “from now on” and treats them as durable learning rather than one-off guidance.
  • Refines into actionable learning: Rewrites the instruction into a clean, typed learning bullet using conventions such as **[convention]**, **[pattern]**, **[anti-pattern]**, **[preference]**, and **[context]**.
  • Confirms and persists: Shows the refined learning for confirmation, then saves it via caliber learn add and stages CALIBER_LEARNINGS.md for the next commit.

Quick Start

Ask the assistant to “remember that for future sessions, always do X” and confirm when it shows the refined learning to save.

Frequently Asked Questions about save-learning

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I save coding preferences so they persist across sessions?

To save coding preferences across sessions, you can use natural remember-style phrasing like "always do X" or "from now on." The system detects these instructions, refines them into typed learning statements, and saves them to a persistent project file for future use.

What is the best way to remember project conventions for future coding tasks?

Remembering project conventions is handled by converting your recurring instructions into typed learning bullets, such as patterns or anti-patterns. After you confirm the refined statement, the learning is added to your project and staged for the next commit.

How do I stop repeating the same workflow instructions every time I start a project?

To stop repeating workflow instructions, you can issue a durable learning command. The system captures phrases like "remember this," rewrites them into an actionable preference, and persists the rule so it is automatically applied in future sessions.

Does saving project learnings require manual git staging?

Saving project learnings does not require manual git staging. When a learning is confirmed, the system automatically adds it to the project file and stages the CALIBER_LEARNINGS.md file for your next commit.

How do I define do and don't rules that apply to all my future work?

Defining do and don't rules requires phrasing your instruction as an enduring preference. The system detects your intent, formats the rule as a convention or anti-pattern, and saves it to ensure the context is retained across time rather than followed just once.