Memorix 1.0.7

Memorix 1.0.7 stabilizes the HTTP control plane, role-aware team flows, dashboard semantics, documentation truthfulness, and Windows build/package reliability.\n\nHighlights:\n- role-aware Team/task/handoff semantics with clearer Project Collaboration UI\n- more honest hooks/integration/runtime signaling and docs alignment\n- background control-plane lifecycle, readiness, and dashboard routing cleanup\n- Windows build determinism and complete release tarball contents\n- fresh verification: tsc, vitest, build, and package all green\n\nAttached asset: npm package tarball (memorix-1.0.7.tgz).

v1.0.6

Added -- Memory Provenance and Layered Retrieval

• Provenance foundation -- Observations now carry sourceDetail (explicit / hook / git-ingest) and valueCategory (core / contextual / ephemeral). All ten write-path call sites annotated. Backward-compatible: old data without new fields parses cleanly.
• Layered disclosure (L1/L2/L3) -- memorix_session_start now separates routing hints (L1), working context (L2), and deep evidence (L3). Session injection scores observations by source and value category so hook noise stays out of working context.
• Evidence retrieval -- memorix_detail and memorix_timeline now surface provenance cues (source badge, evidence basis) so operators can trace why a memory exists and what supports it.
• Verification-aware evidence -- Detail and timeline outputs distinguish direct, summarized, derived, and repository-backed evidence without requiring a full citation framework.
• Citation-lite -- Evidence-bearing surfaces emit lightweight citation hints ([source: git], [verified: repo-backed]) to support "why surfaced" and "what supports this" queries.
• Retrieval tuning -- Source-aware boost treats git-ingest as first-class git evidence for intent-aware ranking. Lightweight provenance tiebreaking for ambiguous retrieval results. L1 routing surfaces active entities as next-hop search guidance.
• Graph routing hints -- Knowledge graph neighborhood is used for lightweight retrieval enrichment and entity-affinity scoring without a full graph rewrite.

Added -- Task-Line Scoping, Secret Safety, and Attribution Hardening

• Task-line scoping -- Search and session context now bias toward the current entity/task-line/subdomain, reducing cross-workstream contamination within a single project bucket.
• Secret safety -- Store-time detection blocks obvious credentials, passwords, and tokens from entering durable memory. Retrieval-time redaction acts as a second safety net for already-stored sensitive data.
• Project attribution hardening -- Write-path consistency checks reduce wrong-bucket writes. memorix_audit_project scans for misattributed observations and reports them with confidence levels.
• Ambiguous-target attribution fix -- Observations stored during ambiguous project context are now flagged rather than silently written to the wrong bucket.

Added -- Retention, Cleanup, and Operator Remediation

• Retention calibration -- Source-aware retention multipliers (hook 0.5x, git-ingest 1.5x) and value-category multipliers (ephemeral 0.5x, core 2.0x) with a 7-day minimum floor. Immunity refined: only critical importance and core valueCategory grant permanent immunity; high-importance types keep long retention but can now decay.
• Retention explainability -- memorix_retention action="stale" shows a full table with per-observation retention explanation (importance, multipliers, effective days, zone, immunity reason).
• Cleanup remediation loop -- memorix_retention (stale/report), memorix_audit_project, and memorix_resolve now form a coherent operator loop. Each output includes structured Suggested IDs: [...] blocks and explicit next-step guidance. memorix_resolve links back to retention report for closed-loop cleanup.

Added -- OpenCode Plugin Improvements

• post_compact event -- New post_compact hook event type. OpenCode's session.compacted event correctly maps to post_compact (was incorrectly mapped to pre_compact). Plugin event handler triggers runHook side-effect on compaction completion.
• Structured compaction prompt -- OpenCode compaction prompt rewritten as a structured continuation format requesting current task, key decisions, active files, blockers, next steps, active entities, and memorix context. No longer promises automatic memorix_store / memorix_session_start invocation during compaction.

Fixed

• #45 OpenCode compaction -- Compaction prompt no longer makes misleading tool-call promises. session.compacted event now fires a real side-effect via runHook. Normalizer mapping corrected to post_compact.
• #46 Dotenv load order -- loadDotenv() now runs before getEmbeddingProvider() in status, doctor, and TUI entry points, fixing "No API key" errors when .env credentials were present.
• #48 Ingest log dedup -- memorix ingest log now deduplicates by commit hash, matching the behavior of ingest commit and TUI batch ingest. Repeated runs skip already-ingested commits.

Stats

• Tests: 1439 passed | 2 skipped (102 files)
• Phases landed: 11 (provenance -> layered disclosure -> evidence -> verification -> citation-lite -> retrieval tuning -> graph routing -> task-line/secret -> attribution -> retention -> cleanup ergonomics)

v1.0.5 - TUI Maturity, Gemini CLI, and Release-Readiness

Highlights

• TUI workbench now feels like a product, with Ink-native /configure, interactive sidebar navigation, unified keyboard behavior, and a cleaner no-project experience.
• Gemini CLI is now a first-class integration target across the TUI, adapters, rules sync, hook normalization, and MCP config generation.
• Silent auto-update is now wired into real runtime entry points instead of living as an unused module.
• memorix_session_start now separates recent handoff context from long-term key project memories.

Stability and correctness

• Fixed embedded dashboard CORS so localhost-only behavior is enforced again.
• Closed the /api/config?project=... startup-project YAML leak.
• Fixed memorix_detail bare numeric IDs so they stay project-safe.
• Fixed topicKey upsert races under concurrent writes.
• Fixed cold-start CLI search by hydrating persisted observations into the index before searching.
• Hardened embedding provider behavior with batch-limit chunking, split-and-retry, and transient 429/5xx retry handling.

Integrations and docs

• Added stale-install detection for the OpenCode plugin so outdated generated plugins can be identified and reinstalled.
• Refreshed the README, Chinese README, architecture/setup docs, and agent-facing docs to match the current cross-agent MCP product shape.
• Added the Agent Operator Playbook, a dedicated AI-facing operator guide for installation, project binding, integrations, hooks, and troubleshooting.

Validation

• npm run build
• npx vitest run
• 82 files, 1099 passed, 2 skipped, 0 failed
• Real smoke coverage across stdio MCP, HTTP control plane, dashboard, TUI workbench, Gemini CLI integration, git-hook ingest, and cold-start CLI search

Known limitation

• Gemini CLI and Antigravity share the official .gemini/* ecosystem. Integrations are independent at the adapter/target level, but hook runtime identity can still behave as "last installer wins" because both share the same official hook config surface.

v1.0.4 — Git Memory, Config Consolidation, Dashboard Control Plane

Highlights

• Git Memory pipeline matured with noise filtering, source-aware retrieval, and git↔reasoning cross-linking.
• Config model consolidated around memorix.yml + .env, with provenance diagnostics in status and dashboard.
• Dashboard upgraded into a Memory Control Plane with Git Memory, Config, Identity Health, and steadier graph behavior.
• Runtime hardening for project detection, dotenv switching, worktree-safe hooks, and cross-project detail.

Validation

• 68/68 test files passing
• 879/879 tests passing
• npm package 1.0.4 published manually with 2FA

See CHANGELOG.md for the full release notes.

v1.0.3 Memory Formation Pipeline

Added

• Memory Formation Pipeline Three-stage pipeline (Extract Resolve Evaluate) runs in shadow mode on every memorix_store call and hooks trigger. Collects quality metrics without affecting storage decisions.
  • Extract: Automatic fact extraction, title normalization, entity resolution, type verification
  • Resolve: 4 resolution actions (new/merge/evolve/discard)
  • Evaluate: Multi-factor knowledge value assessment (core/contextual/ephemeral)
• memorix_formation_metrics tool Query aggregated Formation Pipeline metrics

Stats

• MCP Tools: 23 (+1)
• Tests: 803/803 passing across 60 files (+50 new)
• Hooks: zero modification to existing logic

Full changelog: https://github.com/AVIDS2/memorix/blob/main/CHANGELOG.md

v1.0.2

Fixed

• MCP Server version mismatch Server now reports the correct version from package.json (was hardcoded 0.1.0). Injected at build time via tsup define.
• CI Node.js matrix Removed Node 18 from CI matrix to match engines >=20 in package.json.
• Orama reindex idempotency reindexObservations() now resets the Orama DB before rebuilding, eliminating 'document already exists' errors in multi-session scenarios.
• E2E tests no longer touch real user data Mini-skills E2E tests now use a temporary directory with synthetic observations instead of reading/writing ~/.memorix/data/.

Full Changelog: v1.0.1...v1.0.2

v1.0.1

Fixed

• OpenCode stdout pollution Removed all console.log from the generated OpenCode plugin template. Hooks now run fully silent. (fixes #15)
• OpenCode session_id missing normalizeOpenCode() now reads session_id from the payload instead of hardcoding empty string. Plugin template generates and injects a stable session ID per plugin lifetime. (fixes #14)
• Auto-install hooks scope Hooks are now only auto-installed for IDEs whose project-level config directory already exists (e.g. .cursor/, .windsurf/), preventing unwanted IDE directories from appearing in projects opened with a different IDE.

Added

• MEMORIX_DATA_DIR environment variable Override the default data directory (~/.memorix/data/) by setting MEMORIX_DATA_DIR. Applied consistently across persistence, alias registry, and embedding cache.

Full Changelog: v1.0.0...v1.0.1

v1.0.0 Stable Release

Memorix v1.0.0 — Stable Release

The first stable release of Memorix. All major features are complete. Future versions will iterate based on AI/agent ecosystem evolution.

Highlights

Tool Consolidation (41 → 22 default tools, -46%)

• Team tools: 13 merged into 4 via action parameter (team_manage, team_file_lock, team_task, team_message)
• KG tools: 9 tools now conditional via ~/.memorix/settings.json
• Export/Import: merged into single memorix_transfer tool

Multi-Agent Team Collaboration

• Cross-IDE agent coordination: join/leave, file locks, task boards, messaging
• File-based persistence (team-state.json) — multiple IDE processes share team state
• Verified: Windsurf ↔ Antigravity bidirectional communication

Auto-Cleanup on Startup

• Background retention archiving (expired memories auto-archived)
• Intelligent deduplication:
  • With LLM configured: semantic analysis via any OpenAI-compatible model
  • Without LLM: Jaccard similarity consolidation
• Zero manual maintenance required

Dashboard Team Panel

• Redesigned with Iconify icons (no emojis)
• Agent cards, task lists, message panel, file lock panel

LLM Enhanced Mode

• Compact-on-write: prevents duplicate storage at write time
• Narrative compression: ~27% token reduction
• Search reranking: 60% of queries improved
• Configure via ~/.memorix/config.json or env vars

Full Feature List

• 22 default MCP tools (+9 optional KG tools)
• 10 IDE support: Cursor, Windsurf, Claude Code, Codex, Copilot, Kiro, Antigravity, OpenCode, Trae, Gemini CLI
• Cross-agent memory sharing
• 3-layer progressive disclosure (search → timeline → detail)
• 9 observation types with retention decay
• Mini-skills (permanent, auto-injected knowledge)
• Auto-memory hooks (5 lifecycle hooks)
• Knowledge Graph (MCP official compatible)
• Rules sync across all agents
• Workspace sync (MCP configs, skills migration)
• Web Dashboard (localhost:3210)
• Session management with handoff summaries
• Export/Import (JSON + Markdown)
• 753 tests passing across 56 test files

Install

npm install -g memorix

Full documentation: README