v0.1.1

What's new in 0.1.1

Maintenance release — no changes to recall or write-gate behavior.

• Secret scanner now labels Anthropic keys (sk-ant-…) as anthropic-key instead of openai-key in redaction logs; the openai pattern excludes the ant- prefix so a key yields one finding, not two.
• CI — GitHub Actions runs the test suite on Node 20 + 22 for every push and PR.
• SECURITY.md — private vulnerability-reporting path.

Full Changelog: v0.1.0...v0.1.1

v0.1.0 — Phase A

First public release of automem-synapse — a Claude Code plugin that makes AutoMem automatic and safe.

• Automatic recall at session start and before every turn (via the mcp-automem sidecar).
• Fail-closed write gate on every memory write: normalize → secret-scan → policy → dedupe. Secrets blocked, duplicates flagged, sensitive categories confirmed.
• /automem-status + /automem-recall commands, usage skill, JSONL observability.
• Zero runtime dependencies; Node ≥ 20. Codex-APPROVED; 61 offline tests + live smoke; first-live-use verified.

See the README for install + configuration.