Reproducible methodology

Fix the question. Blind the provider. Publish the limits.

The harness is designed to prevent the website from becoming stronger than the evidence.

Experimental contract

One controlled question per track.

This publication asks whether a governance layer removes invalid exposure without blocking valid evidence.

6 cases

Public synthetic governance fixture.

12 documents

Valid and deliberately invalid memory states.

Top-1 / K=4

Provider returns four candidates; scorer evaluates top-1.

5 repeats

Sequential, distinct clean-state processes.

Fairness controls

Provider retrieval stays provider retrieval.

Blinded provider input

The provider receives redacted documents and queries, not policy labels or the full evaluation fixture. Gold governance labels stay in the policy replay and scorer processes.

Paired comparison

The same ranked candidates feed baseline and governed arms. Vault's guard filters against canonical state; it does not silently swap in a different provider.

Clean-state identity

Each repeat uses a distinct provider store or namespace, verifies index completion, and binds execution evidence to raw runs.

Honest missingness

Unavailable provider cost and unmeasured integrations remain unavailable. Missing data is never rendered as zero.

Publication gates

Fail closed before a number becomes a headline.

GatePurposePublished bundle
Checksums and unique artifactsDetect mutation or accidental reusePassed
Clean source chainBind results to committed benchmark codePassed
Blinded provider inputKeep evaluation labels away from retrievalPassed
Provider integrity and preflightVerify native retrieval actually ranPassed
Five distinct clean-state repeatsCheck deterministic reproducibilityPassed
Zero indexing failuresAvoid scoring partial corporaPassed
Verify independently

Check all 36 evidence files with one command.

python scripts/verify_publication_bundle.py \
  benchmarks/results/vaultgovbench-retrieval-v0.1/89b9156

The verifier fails closed on changed or unlisted files, unsafe paths, invalid indexes, or non-publishable summaries. A pass verifies artifact integrity—not an independent provider rerun.

Inspect the verifier source →

Source identity

fixture sha256:
017c0629a561…c3b0b5f

provider input sha256:
ddd3a0799c46…41c6f4fc4

source revision:
89b9156f501b…eb159246b4b1db

Rebuild the public view

Generate the website from the verified evidence.

python scripts/build_benchmark_site_data.py
python tests/test_benchmark_site.py

Run the full experiment program →

Limitations

What this evidence does not prove.

Not end-to-end QA

It measures retrieval governance, not final answer generation or agent task success.

Not a universal claim

One published external provider does not prove improvement for every memory engine.

Not production scale

Six synthetic cases and twelve documents do not measure large-corpus throughput or multi-agent contention.

Next evidence program

From contract proof to category proof.

Clean AgentMemory and Letta publications; neutral LoCoMo/LongMemEval retrieval tracks; scale curves; concurrent multi-agent operations; failure injection; native extraction cost and quality; third-party reproduction.