v1.4.0 — Faulkner extractor chaining + sync hardening

What's New in v1.4.0

Nightly automation got substantially more capable. The sync-and-index.sh script can now chain into the Faulkner-DB relationship extractor after indexing completes, so if you use both projects together your knowledge graph stays fresh on the same cron that indexes your conversations — no second scheduler entry needed.

Highlights

• Optional Faulkner-DB extractor chaining. After /index/trigger completes (polled via the new /index/status endpoint, 15 min ceiling), the script can fire relationship_extractor.py --incremental against your FalkorDB instance. A 20-hour run-gate on reports/extraction_state.json mtime prevents duplicate work. The step is wrapped in set +e and skips gracefully if the faulkner-db venv or state file isn't present — the sync works fine without it. Opt-in via the env vars documented in the README.

• ~/.claude/docs/ rsync. The script now also syncs your local ~/.claude/docs/ directory to the remote host if it exists. This is a landing pad for session-handoff files and any other persistent notes you want searchable across sessions via a downstream doc-search layer. Safe no-op if the directory doesn't exist.

• Sanitized public defaults. All script defaults now target localhost and use changeme placeholder passwords. Production config lives in env vars (GENESIS_REMOTE_HOST, FAULKNER_LLM_ENDPOINT, FALKORDB_*), never in committed source. Safe to clone and adapt without stripping credentials.

New API Endpoint

GET /index/status — poll the current indexing job state (no_job / running / complete / failed). Used by the sync script's polling loop; also handy for monitoring.

Internal Changes

• daemon/mkg_client.py: this file is an internal helper used only when Agent Genesis is deployed alongside a private local-LLM analysis gateway. It was previously a stub that returned a placeholder string; it now makes a real HTTP call via AGENT_GENESIS_LLM_ENDPOINT so that helper works when wired up. Not a feature of the public project — you don't need to configure it, and standard Agent Genesis usage (/search, /index/*, stats, etc.) doesn't touch it.

Upgrade Notes

• No schema or data migration needed.
• If you were running the previous sync-and-index.sh with an unset GENESIS_REMOTE_HOST, the script exits cleanly when it's unset; set it explicitly to enable remote sync.
• If you want to opt into the Faulkner extractor chaining, check out faulkner-db v1.6.0 alongside this release — they're designed to be used together.

Full Changelog

See commit 74170bf for the complete diff.