remember

Routes user memory requests to the correct persistence destination via a decision tree.

3|1|Updated Dec 9, 2025
One-click install
npx skills add https://github.com/isizono/calm --skill remember-isizono
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: remember
Source: https://github.com/isizono/calm/tree/main/skills/remember
Command: npx skills add https://github.com/isizono/calm --skill remember-isizono

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When a user asks an AI assistant to remember something, the information could belong in many places—feedback entries, habits, rules files, tag-notes, pins, auto-memory, or precedent decisions. This Skill provides a mandatory routing flow that determines the correct storage destination so memories are actually retrievable at the right moment instead of being written to the wrong place and lost. ## Core Features & Use Cases - Destination Decision Tree: Classifies each memory request by trigger, type (unconditional rule, conditional judgment criterion, or fact/tip), and risk to select the correct storage target. - Recurrence Handling: Detects repeated corrections via deja vu triggers, records stumble notes, and escalates rules to stronger enforcement (block entries, rules, habits) after repeated failures. - Precedent Format Support: Ships a reference specification for writing conditional judgment criteria as structured precedent decisions with rejected alternatives, scope, and verification anchors. - Use Case: A user says "remember to never push directly to main." The Skill checks existing records, determines this is an unconditional rule with irreversible risk, and routes it to a block-level feedback entry or rules file after user approval. ## Quick Start When I ask you to remember or save something, use the remember skill to decide where it should be stored.

Frequently Asked Questions about remember

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

FAQPage Schema
How do I make an AI assistant remember a rule permanently?▼

Tell the assistant to remember the rule, and the remember skill classifies it by trigger, type, and risk to pick a destination. Unconditional rules with high violation cost go to habits or rules files with approval, while low-priority tips go to auto-memory.

What happens when the same correction is repeated to an AI?▼

The skill detects recurrence via deja vu triggers and checks existing records first. It records a stumble note on the feedback entry, and after three recurrences escalates the rule to a stronger destination like a block entry, rules file, or always-on habit.

Where should conditional judgment rules be stored in CALM?▼

Conditional judgment criteria ("if X then Y is allowed") are stored as precedent decisions using add_decisions with scope-in and scope-out clauses. A summary and reference to the precedent is also placed at the trigger location such as tag-notes or pins.

Can the remember skill edit or delete existing memories?▼

No, the remember skill only handles new saves. Correcting, rewriting, or deleting records that are outdated or wrong is delegated to the separate forget skill, which the decision tree routes to when stale records are found.

Why do some saved rules never surface during AI sessions?▼

Pull-based destinations like auto-memory and intelligently-tier habits are often not retrieved in practice. Rules with irreversible risk must be placed on push paths such as block entries, pins, tag-notes, or always-tier habits to guarantee delivery.