Skills and memory

Skills

A skill (ai.skill) is a reusable instruction pack the assistant can load mid-task — similar to progressive disclosure in modern coding agents.

Menu: AI ‣ Configuration ‣ Behavior ‣ Skills

Fields

Field

Purpose

Name

Short UI / catalogue label

Code

Technical id for use_skill (^[a-z][a-z0-9_]*$)

Description

When to use (English; shown to the model)

Body

Full instructions when active (English)

Activation mode

On demand (catalogue + use_skill / keywords) or Always (every turn — use sparingly)

Trigger keywords

Comma-separated keywords on the latest user message for auto load

Active / sequence

Visibility and ordering

Provenance

Factory / custom (and factory-modified)

Runtime behaviour

  • CRUD is AI Administrator only.

  • At runtime every AI user sees the same active skills (no per-user skill toggle).

  • Tools list_skills / use_skill require capability ask.

  • Bounds limit catalogue size, always-on count and body length so one skill cannot dominate the prompt.

Factory skills

The module ships factory skills (XML data, noupdate). Examples include operational playbooks for vendor bills, bank reconciliation patterns, follow-ups and localisation-aware product fields — grounded to eYssen stacks where applicable.

  • Sync factory skills admin action refreshes pristine factory rows from code on upgrade.

  • If you edit a factory skill, it becomes factory-modified and is not auto-overwritten; you can reset to factory when desired.

Writing a good skill

  • State triggers clearly in the description.

  • Give ordered steps and which tools to call first.

  • Say what not to invent (accounts, taxes, partners).

  • Keep always-on skills short; put long playbooks on demand.

Memory

Memory (ai.memory) stores durable facts the assistant may recall across conversations.

Menu: AI ‣ Memory ‣ Memories and Memory Graph.

Types

  • Working — short-lived task context

  • Episodic — what happened in a session

  • Semantic — durable facts / preferences

  • Archive — cold storage

Scopes

  • User (default private)

  • Company / System — admin (or approved promotion)

  • Agent / Run — agent-centric context

Users may create/write user, agent and run scopes; company/system need admin or promotion approval.

Tools

Under capability ask: memory_search, memory_save, memory_forget, memory_link. Prompt injection of memories is size-bounded.

Good practice

  • Save stable preferences (“user wants HU date format”) as semantic/user.

  • Do not store passwords, API keys or full personal ID documents in memory.

  • Review company-scope memories periodically; treat them as shared configuration.

  • Use forget when a fact is wrong — do not pile contradictions.

Skills vs memory

Skills

Memory

How to do a job (playbooks)

What is true for this user/org (facts, preferences)

Admin-authored configuration

User/agent/runtime authored

Versioned factory possible

Operational data