# zettelforge > Agentic memory for CTI in Python — STIX knowledge graphs, threat-actor alias resolution, offline-first RAG, MCP server for Claude Code and LangChain agents - **URL**: https://www.freshcrate.ai/projects/zettelforge - **Author**: rolandpg - **Category**: MCP Servers - **Latest version**: `packages-v0.1.0` (2026-05-26) - **License**: MIT - **Source**: https://github.com/rolandpg/zettelforge - **Homepage**: https://threatrecall.ai - **Language**: Python - **GitHub**: 25 stars, 4 forks - **Registry**: github (`rolandpg/zettelforge`) - **Tags**: `agentic-memory`, `ai-agent`, `claude-code`, `cti`, `cybersecurity`, `knowledge-graph`, `langchain`, `llm`, `python` ## Description Agentic memory for CTI in Python — STIX knowledge graphs, threat-actor alias resolution, offline-first RAG, MCP server for Claude Code and LangChain agents ## Recent releases | Version | Date | Urgency | Changes | | --- | --- | --- | --- | | `packages-v0.1.0` | 2026-05-26 | High | Release tag for packages v0.1.0. | | `v2.7.0` | 2026-05-26 | High | ## ZettelForge v2.7.0 Security-focused release adding write-time MemSAD memory defenses with audit, block, and quarantine modes; governance.memory_defense configuration; RFC-017 threat model and rollout plan; telemetry actor/caller compatibility; and regression coverage. Validation before release: - Local full suite: 742 passed, 13 skipped - GitHub protected checks: CI, Snyk Security, lint, pip-audit, governance, Python 3.12/3.13 tests, package build - Package build and twine check passed for | | `v2.6.2` | 2026-04-27 | High | UI/UX release. Fixes the `/config` page so the Apply button actually works and surfaces enum-style settings as dropdowns instead of free-text inputs. No data migration. No config changes. No API contract changes. ## Fixed - **`/config` "Save Changes" button is no longer dead.** The Quick Settings panel called `saveConfigForm()` and `reloadConfig()` — neither function was defined anywhere, so the button silently no-op'd and the panel rendered "Loading schema..." forever. Replaced with a real fo | | `v2.4.1` | 2026-04-24 | High | **Operational telemetry (RFC-007), SQLite backend fixes, and TypeDB authentication hardening.** ## Added - **Operational telemetry (RFC-007)** ([#85](https://github.com/rolandpg/zettelforge/pull/85)) — new per-query metrics stream at `~/.amem/telemetry/telemetry_YYYY-MM-DD.jsonl` when `ZETTELFORGE_LOG_LEVEL=DEBUG`. Five shipped components: - **`TelemetryCollector`** class with `start_query` / `log_recall` / `log_synthesis` / `log_feedback` / `auto_feedback_from_synthesis`. INFO/DEBUG-gated f | | `v2.4.0` | 2026-04-19 | High | **Detection-rules-as-memory, MCP Registry publication, SQLite concurrency hardening, test-suite hygiene, and brand/docs polish.** ## Added - **Sigma + YARA as first-class memory entities** with LLM rule explainer ([#70](https://github.com/rolandpg/zettelforge/pull/70)) - **Detection Rules as Memory** README section ([#74](https://github.com/rolandpg/zettelforge/pull/74)) - **MCP Registry publication infra** — \`server.json\` and PyPI \`mcp-name\` tag so ZettelForge can be published to [registry | | `v2.3.0` | 2026-04-17 | High | ## [2.3.0] - 2026-04-17 Pluggable LLM provider infrastructure (RFC-002 Phase 1), MCP server as a first-class Python module, PyPI discoverability refresh, SEO foundations across the docs site, and a full docs-vs-code reconciliation. All additions are backward-compatible; no existing API changes. Supersedes the never-tagged 2.2.1 metadata patch — its PyPI classifier / keyword / image-URL changes are folded in below. ### Added - **Pluggable LLM provider infrastructure (RFC-002 Phase 1)** — new | | `v2.2.0` | 2026-04-16 | High | ## What's New **SQLite is now the default storage backend.** Zero-config, ACID guarantees, WAL mode. LanceDB stays for vector search. No Docker, no TypeDB, no external services required. ### Highlights - **SQLite default backend** — `StorageBackend` ABC (33 methods), `SQLiteBackend` (700+ lines), backend factory with auto-detection. Migration script for existing JSONL data. - **Causal chain retrieval** — Fixed critical bug where all LLM-extracted causal edges were invisible. Added reverse tra | | `v2.1.1` | 2026-04-15 | High | ## What's New ### Production Blockers Fixed - O(n) supersession scan replaced with O(E) entity index lookup - File locking on all JSONL writes (fcntl) - SQL injection sanitized (5 sites in VectorMemory) - LanceDB dedup guard prevents ghost rows - Entity index invalidated on supersession ### Performance - Dual-stream write: remember() returns in ~45ms (was 700-3500ms) - Async entity index saves (5-second write-behind) - KG JSONL compaction added to rebuild_index.py - Batched access_count persis | | `v2.1.0` | 2026-04-12 | Medium |