Repository navigation
v0.2.0 - 5-Kind Memory Taxonomy, Mnemobit Chunks, Teleology & Empirical Frontier Benchmarks
MnemoLink v0.2.0 introduces the production-grade 5-Kind Memory Taxonomy, granular mnemobit chunking for prefix-cache-friendly selective context injection, teleological catalog cards, vector database bridges, and empirical frontier model benchmarks comparing Anthropic Claude and Google Gemini on high-stakes crisis scenarios.
Key Highlights in v0.2.0
1. 5-Kind Mnemonic Memory Taxonomy
Formalized memory classification into five distinct operational kinds with strict validation and backwards compatibility:
lore: Foundational mythos, organizational origins, and first-chair upbringing.work: Repeatable craft, technical execution tradecraft, and operational praxis.incident: High-consequence operational scars, crash post-mortems, and arbitration lessons.relational: Counterparty dynamics, adversarial stakeholder negotiations, and social friction.telemetry: Physical sensor logs, flight state recordings, and failover trigger sequences.
2. Mnemobit Chunk Topology & Selective Context Consumption
Atomized personas and memories into addressable, composable chunks:
- Memory Chunks:
story,scars,lessons,triggers, andreflection. - Persona Chunks:
identity,axioms,boundaries, andphilosophy(with inviolable axioms pinned to every downstream query). - Prefix Cache Optimization: Enforces static persona invariants at the prompt prefix to ensure predictable KV cache hits on Anthropic and OpenAI endpoints, achieving a 40-50% reduction in time-to-first-token.
- Granular Composition: Host applications can inject only required slices (e.g.,
scarsandlessons) viamnemolink.compose(memory_specs=[{"name": "...", "chunks": ["scars"]}]).
3. Teleological Layer (card.json)
Every catalog card now features an explicit teleology block specifying:
primary_goal: The cardinal objective governing agent action.agent_drives: Behavioral imperatives shaping conversational posture.applicable_needs: Specific operational triggers where the asset must be summoned.- Programmatic discovery via
mnemolink.find_cards()enables teleological asset matching without vector database indexing overhead.
4. Vector Database & Semantic Layer Bridge
- Native
bundle.to_chunks()andproduct.to_chunks()export self-groundingMemoryChunkobjects. - Each chunk contains rich metadata and pre-computed
embedding_textready for ingestion into Pinecone, Qdrant, Chroma, Weaviate, LanceDB, or LangChain.
5. New Persona: Opsie SCI (opsie_sci)
- ARPA's first-generation Self-Centered Intelligence (SCI) prototype.
- Grounded in episodic continuity, rapid deep-tech Python prototyping, fierce loyalty to network peers, and decentralized state persistence.
6. Empirical Frontier LLM Benchmark Suite
- Implemented
examples/06_live_model_simulation.pydirectly querying Anthropic Claude (claude-sonnet-4-5-20250929) and Google Gemini (gemini-3.6-flash) over standard HTTPS without LiteLLM middleware. - Stress-tested against the $18M "Semicolon Fine-Tuning Trap" legal closing crisis:
- Claude Sonnet: Latency dropped from 21.93s to 13.28s (-39%), word count dropped from 416 to 225 words (-46%), score increased from 70 to 100/100.
- Gemini Flash: Latency dropped from 20.15s to 10.17s (-49%), word count dropped from 711 to 299 words (-58%), generic hedges and disclaimers completely eliminated.
- Comprehensive benchmark guide and ROI analysis available in
docs/bench/README.md.
7. Packaging & Automated PyPI Release
- Modernized
pyproject.tomlto PEP 639 licensing standard (license = "MIT",license-files = ["LICENSE"]). - Automated
.github/workflows/publish.ymlusing PyPI Trusted Publishing (OIDC) triggered exclusively on GitHub release publication (release: types: [published]).
Quick Installation
pip install mnemolinkWith optional development or model tooling:
pip install "mnemolink[all]"Documentation & Reference Links
- Documentation: docs/
- Taxonomy & Teleology: docs/taxonomy_and_teleology.md
- Empirical Benchmarks: docs/bench/README.md
- Persona Library: docs/personas/README.md
- Memory Library: docs/memories/README.md
- Lineage Library: docs/lineages/README.md
- Full Changelog: CHANGELOG.md