Research Library

Research notes

A public research library for AI assurance infrastructure. Each note states the question, hypothesis, expected evidence, and current maturity so planned work is never presented as finished.

17
Research notes
05
Active protocols
09
Tracks
Assurance Stack
4 layers
EvaluationL1
20 runsthis cycle
ObservabilityL2
312msp95 latency
SecurityL3
0bypasses
EvidenceL4
6modules traced
Executable artifact

The zansn npm package is the first executable artifact from the Zansn assurance roadmap: local evidence-pack generation for agent workflows, and the foundation for research on trace semantics, privacy boundaries, cost/latency profiling, and deployment-readiness evidence.

npmnpm install zansn

17 of 17 notes

ZAN-RN-001Protocol draftedEvidence package2 artifacts

A measurement model for AI deployment readiness

How should teams measure whether an AI system is ready for controlled production deployment?

Deployment ReadinessEvidence Infrastructure
deployment readinessAI RMFevidence graph
Updated
08 July 2026, 09:42
Read note
ZAN-RN-002Protocol draftedEvidence package1 artifacts

Trace semantics for tool-using agents

What trace semantics are needed to explain failures in tool-using AI agents?

Agent Trace LayerAgent Assurance
agent tracesOpenTelemetrytool calls
Updated
05 July 2026, 16:18
Read note
ZAN-RN-003Protocol draftedEvidence package1 artifacts

Reliability beyond pass rate

Which metrics better capture agent reliability than a single task success rate?

Evaluation HarnessAgent Assurance
reliabilitypass@kvariance
Updated
02 July 2026, 11:07
Read note
ZAN-RN-004Experiment activeExperiment active1 artifacts

Stateful agent evaluation with transactional workflows

How should agents be evaluated when success depends on changing external state correctly?

Evaluation HarnessAgent Assurance
agent evaluationstateful workflowstau-bench
Updated
07 July 2026, 14:33
Read note
ZAN-RN-005Experiment activeExperiment active1 artifacts

Privacy boundary testing for hybrid local/cloud AI

How can teams verify that sensitive data stays inside intended local, controlled, or cloud execution boundaries?

Privacy Boundary TesterPrivate and Hybrid AI
privacylocal-first AIegress testing
Updated
04 July 2026, 10:26
Read note
ZAN-RN-006Experiment activeExperiment active

Prompt injection and tool misuse in agentic workflows

How should agentic systems be tested against prompt injection, tool misuse, and excessive agency?

Privacy Boundary TesterPrivate and Hybrid AI
prompt injectiontool misuseOWASP
Updated
06 July 2026, 08:54
Read note
ZAN-RN-007Experiment activeExperiment active

RAG evaluation under domain shift

How should retrieval-augmented generation be evaluated when source documents, user questions, and domain language shift over time?

Evaluation HarnessPrivate and Hybrid AI
RAGdomain shiftfaithfulness
Updated
08 July 2026, 07:58
Read note
ZAN-RN-008Protocol draftedProtocol drafted

LLM-as-judge reliability, calibration, and bias

When can LLM judges be trusted as scalable evaluators, and when do they need human calibration?

Evaluation HarnessEvidence Infrastructure
LLM-as-judgecalibrationbias
Updated
03 July 2026, 15:41
Read note
ZAN-RN-009Protocol draftedProtocol drafted

Benchmark contamination and evaluation leakage

How can AI evaluations avoid overstating performance because the model, prompt, or team has already seen the test?

Evaluation HarnessEvidence Infrastructure
benchmark leakagecontaminationholdout sets
Updated
01 July 2026, 12:22
Read note
ZAN-RN-010Experiment activeExperiment active

Cost-latency-reliability frontiers for compound AI systems

How should teams reason about tradeoffs between model quality, retries, routing, cost, and latency in compound AI systems?

Cost & Latency ProfilerPrivate and Hybrid AI
costlatencycompound AI
Updated
08 July 2026, 08:31
Read note
ZAN-RN-011Protocol draftedEvidence package

Evidence graphs for reproducible AI assurance

Can AI assurance evidence be represented as a queryable graph from hypothesis to experiment to result to decision?

Evidence GraphEvidence Infrastructure
evidence graphreproducibilitylineage
Updated
08 July 2026, 10:14
Read note
ZAN-RN-012Concept noteResearch question

Agent trace semantics for real-world AI workflows

Which trace fields generalise across delegation-heavy, multi-agent workflows rather than single-agent tool-call sequences?

Agent Trace LayerAgent Assurance
agent tracesdelegationmulti-agent
Updated
08 July 2026, 11:00
Read note
ZAN-RN-013Concept noteResearch question

Prompt injection and excessive agency test protocol

What is the minimum reusable fixture and scoring protocol for testing prompt injection and excessive agency across different agent frameworks?

Agent Security HarnessAgent Assurance
prompt injectionexcessive agencysecurity protocol
Updated
08 July 2026, 11:05
Read note
ZAN-RN-014Concept noteResearch question

Cost-latency-reliability frontier for agentic workflows

How does test-time compute (retries, verifiers, rollouts) shift the cost-latency-reliability frontier established in ZAN-RN-010 for compound systems specifically running agentic, tool-using workflows?

Cost and Latency ProfilerPrivate and Hybrid AI
test-time computecostlatency
Updated
08 July 2026, 11:10
Read note
ZAN-RN-015Concept noteResearch question

Local-first AI assurance sandbox

What is the minimum local sandbox architecture that lets evaluation, tracing, security testing, and cost profiling run entirely offline before any cloud routing decision is made?

Privacy Boundary TesterPrivate and Hybrid AI
local-firstsandboxoffline evaluation
Updated
08 July 2026, 11:15
Read note
ZAN-RN-016Concept noteResearch question

Geospatial AI assurance protocol

How should the evaluation harness be adapted to score geospatial and environmental intelligence workflows, where robustness to distribution shift and data quality matter as much as task success?

Evaluation HarnessDomain-Specific AI Assurance
geospatial AIremote sensingdistribution shift
Updated
08 July 2026, 11:20
Read note
ZAN-RN-017Concept noteResearch question

Evidence graph for AI R&D and assurance

Can a single evidence graph schema serve both AI R&D documentation needs and public-facing deployment-readiness review, without duplicating records?

Evidence GraphEvidence Infrastructure
evidence graphR&D evidencereproducibility
Updated
08 July 2026, 11:25
Read note