Skip to content

v0.5.1 — delegated outcomes now reach memory

Choose a tag to compare

@aspotton aspotton released this 16 Sep 18:24
· 85 commits to main since this release
a0bbe27
  • Learn payloads can carry a capped rendering of the current turn's tool results (1200 characters per message, 4000 total), so outcomes reported by delegated sub-agents are learned once from the parent session.
  • The extraction prompt gains a matching TOOL RESULTS section only when tool results are present; tool-less requests keep byte-identical payloads and prompts.
  • Sub-agent session traffic itself remains gated off by learning.skip_subsessions; no config or schema changes.