mind-mem v4.2.4

Full Changelog: v4.2.3...v4.2.4

mind-mem v4.2.3

Full Changelog: v4.2.2...v4.2.3

mind-mem v4.2.2

Full Changelog: v4.2.1...v4.2.2

mind-mem v4.2.1

Full Changelog: v4.2.0...v4.2.1

mind-mem v4.2.0

Full Changelog: v4.1.2...v4.2.0

mind-mem v4.1.2

Patch over v4.1.1: send_message now logs a best-effort index-rebuild failure at debug level instead of swallowing it with a bare pass (closes Bandit B110 code-scanning alert). No API change.

Full Changelog: v4.1.1...v4.1.2

v4.1.1 — docs/badges/PyPI alignment

Documentation-only patch over v4.1.0: README version/badges aligned, agent-to-agent comm + recall workspace-resolution surfaced, PyPI page re-rendered. No functional changes vs v4.1.0.

v4.1.0 — recall workspace resolution + agent-to-agent comm

Fixed

• recall CLI silently read the wrong store. mind-mem-recall defaulted to an empty local index and printed No results found. even when the configured workspace held matches. Recall now resolves the workspace (--workspace > $MIND_MEM_WORKSPACE > nearest mind-mem.json > cwd) and warns loudly on an empty/unconfigured store instead of a silent miss.

Added

• Agent-to-agent comm (agent_messaging): the sanctioned in-fleet channel for agents to send/receive messages through the governed memory store, surfaced as mm send/mm inbox CLI verbs + an MCP tool. See docs/agent-comm.md.

Full suite: 5559 passed.

mind-mem v4.0.18

Full Changelog: v4.0.17...v4.0.18

Full Changelog: v4.0.17...v4.0.18

mind-mem v4.0.17

Full Changelog: v4.0.16...v4.0.17