nethical6
Featured
conversation-steganography
A tool that hides encrypted secret messages inside innocent-looking AI-generated chat text for private communication over any messaging platform.
machine-genome
Machine Genome is an open identity and provenance protocol for models, agents, datasets, and related artifacts, providing content-addressed, cryptographically signed lineage records without a blockchain.
Stars
274
7-day growth
No data
Forks
7
Open issues
0
License
Apache-2.0
Last updated
2026-07-20
Why it is worth attention
It offers verifiable, append-only provenance graphs for AI components with standard cryptographic proofs, enabling trust in composite agent systems without relying on any single vendor or blockchain.
Who it is for
- Developers building modular AI agents and pipelines
- MLOps engineers managing model and dataset lineages
- Security auditors requiring cryptographic evidence of artifact provenance
- Researchers studying reproducibility and trust in AI systems
Use cases
- Tracking the exact composition of an AI agent from model weights, prompts, and tools
- Verifying the provenance of a dataset used in training a model
- Auditing the history of amendments and attestations to a machine-learning artifact
- Establishing stable, content-addressed identifiers for AI components across registries
Strengths
- Content-addressed IDs (SHA-256 multihash) ensure immutable references to genesis records
- Append-only lineage with typed parent edges preserves history without mutation
- W3C eddsa-jcs-2022 cryptographic proofs bind records to DID controllers
- No blockchain dependency; works with standard infrastructure and offline key resolution
Considerations
- Protocol is experimental (MGS 0.1.1) and not yet stable for production use
- Does not prove inherited capability, consciousness, ownership, or legal status—only signer and bytes
- Current DID resolver is limited to Ed25519 did:key; DNS namespace control requires external challenge
README quick start
Quick start
Description
Machine Genome is an open identity and provenance protocol for models, agents, harnesses, datasets, and the artifacts that connect them.
Related repositories
Similar projects matched by category, topics, and programming language.
programmersd21
Featured
flow
A terminal dashboard that shows your internet throughput in real time, with a deliberately minimal interface designed to be understood within one second.
vshulcz
deja-vu
deja is a zero-dependency binary that indexes local agent conversation histories from multiple AI coding tools into a searchable, private memory layer with MCP recall, redaction, and sync capabilities.