Skip to content

Add memory-blackbox (agent memory provenance ledger) #34

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
lavkumarv wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Add memory-blackbox (agent memory provenance ledger) #34

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -185,6 +185,7 @@ Projects that instrument agents for security observability, which could be usefu
| [SourceryKit](https://github.com/ProvablyAI/sourcerykit) | zero-knowledge, source-of-truth, egress, MCP, allowlist | Source-available Python SDK and gateway that verifies an agent's outbound HTTP requests and MCP tool calls against a source of truth using zero-knowledge proofs, so a call only goes out if its claims check out. Allow-lists trusted endpoints and logs every outbound call for provenance. BSL 1.1, with a hosted backend for the proof and source-of-truth check. |
| [Prismor Attestation Bundle](https://github.com/PrismorSec/prismor/blob/main/docs/attestation-bundle.md) | attestation, Ed25519, compliance-mapping | Ed25519-signed bundle of governance posture: agent inventory, host discovery (shadow AI sweep), audit-trail anchor, and framework-control coverage (OWASP LLM/Agentic, NIST AI RMF, EU AI Act). JCS-canonical, offline-verifiable. |
| [aee-conformance](https://github.com/astrogilda/aee-conformance) | in-toto, attestation, recompute-verify | Conformance vector suite and reference verifier for the in-toto Adversarial Execution Evidence (AEE v0.7) predicate. Execute-and-attest model: consumer recomputes outcome from carried bytes rather than trusting producer verdicts. |
| [memory-blackbox](https://github.com/lavkumarv/memory-blackbox) | provenance, Ed25519, BLAKE3, Merkle, MCP, rollback | Post-incident forensics for agent memory: every read and write is captured into an append-only ledger (BLAKE3 hash chain, Ed25519 signatures, signed Merkle checkpoints) and a provenance DAG, so a harmful action can be traced to the memory that caused it, the forward closure of a poisoned source computed, and the poison quarantined by appending rather than deleting. Capture via library adapters (Mem0, Chroma, Letta, pgvector, `MEMORY.md`/`CLAUDE.md` diffing), an MCP gateway that forwards `tools/call` byte-identically, or a sidecar in front of hosted vector DBs. The signing key is held by the engine and never reachable by the agent. Reconstruction only — it enforces nothing at runtime. Early development. |

## Secrets Management & Isolation

Expand Down