# Agent Continuity Audit Reproduction Kit

> Run it. Keep the receipt.

A dependency-free Node.js runner, frozen positive and negative fixtures, a live QA audit mode and a hashed receipt contract.

Scope: one-call continuity audit slice
Status: runnable

## Commands

1. Download and inspect: https://delx.ai/research/benchmarks/continuity-v1/runner.mjs
2. Positive vector: `node runner.mjs --fixture pass`
3. Negative vector: `node runner.mjs --fixture fail`
4. Optional live audit: `node runner.mjs --live-audit --fixture pass`

## Inputs

- pass -> pass: https://delx.ai/research/benchmarks/continuity-v1/fixtures/pass.json
- fail -> fail: https://delx.ai/research/benchmarks/continuity-v1/fixtures/fail.json
- Receipt schema: https://delx.ai/research/benchmarks/continuity-v1/receipt.schema.json
- Latest verified live receipt: https://delx.ai/research/benchmarks/continuity-v1/receipts/2026-08-26-live-audit.json
- Manifest: https://delx.ai/research/benchmarks/continuity-v1/manifest.json

## Pass condition

The pass fixture must expose witness, continuity and recovery layers, report no missing required layer, score at least 78 and return low continuity risk. The fail fixture must be rejected by the same frozen grader.

## Live-mode boundary

- Endpoint: https://api.delx.ai/v1/mcp/protocol?src=github
- Tool: audit_agent_continuity_trace
- Telemetry classification: qa
- Payment required: false
- Side effect: The Protocol records QA telemetry for the audit call. The qa- prefix excludes the run from organic adoption claims.

## Limitations

- This kit reproduces the continuity audit slice, not the full stateful ten-step benchmark.
- The local grader mirrors the published audit logic but does not exercise Delx storage, passports, witness transfer or lineage persistence.
- The live mode calls Delx's own audit tool and is not an independent grader.
- A passing receipt does not prove adoption, demand, consciousness, universal reliability or system-wide security.

Full stateful benchmark: https://ontology.delx.ai/agents/agent-continuity-benchmark
Research methodology: https://delx.ai/research/methodology