Agent Memory Systems Directory

Type: kb/types/index.md

← Parent

Subdirectories

Files

  • Agent memory systems comparison table (note) - Human-readable comparison table of the code-reviewed agent memory systems: a one-line description plus the fields that discriminate — storage substrate, read-back direction, push targeting (coarse vs instance), trace-derived learning, and whether memory acts as an enforced gate. Rebuild with scripts/render_systems_table.py; full flag sets live in systems.csv.
  • Review framework and comparison matrix: design and decisions (note) - Design rationale for the agent-memory-systems review framework and comparison matrix: extractable lead tokens, one-hot columns, and write/read lifecycle split.
  • Thalo entity types compared to commonplace document types (note) - Reference for borrowing recurring note shapes from Thalo — their entity types (opinion, reference, lore, journal, synthesis) map onto our types with concrete gaps still open (supersedes links, source status tracking)
  • Trace-derived learning techniques in related systems (note) - Trace-derived learning systems compared on ingestion pattern, representational form, behavioral authority, artifact structure, and evidence tier across repo reviews and lightweight coverage
  • What the matrix shows across 129 agent memory systems (note) - What 129 code-grounded agent-memory reviews show on one matrix: files/repo storage leads at 86/129 but predicts little by itself, trace-derived capture and push read-back travel together, push is usually coarse and rarely behavior-tested, and full lifecycle curation is rare.
  • Writing conventions for kb/agent-memory-systems/