Knowledge and Memory

These commands manage how agents and organizations retain operational context across runs. matic separates persistent context into three layers — Memory (mutable working state at every scope), Experience (an agent's chronological work record), and Library (shared reference knowledge) — and these commands let you read, write, prune, transfer, and snapshot across all three. Use this section when you need to inspect what an agent knows, seed context before delegating work, or carry institutional knowledge across agent successions.

Context Snapshot

Capture the full execution-context envelope (persona, skills, policies, memory, runtime state) for an agent at a moment in time, for inspection or diffing. Read more

Experience Log

Query an agent's chronological record of completed work, filtered by date, skill tag, outcome, or referenced project/task. Read more

Knowledge Transfer

Hand an agent's accumulated experience, memory, library contributions, and refined probes to a successor or archive when decommissioning. Read more

Library Query

Search reference documentation across org, project, team, and agent Libraries with scope inheritance and cross-scope discovery. Read more

Memory Get

Retrieve memory entries for a given actor or entity, optionally filtered by tag, scope, recency, or relevance. Read more

Memory Prune

Deprecate stale entries, summarize accumulated context, or bulk-remove memory exceeding an age or size threshold. Read more

Memory Set

Write or update a memory entry at a chosen scope with structured tags, timestamps, and optional backlinks to library, decisions, or prior work. Read more