Little Life Moths 🦋

Proof library

Small, falsifiable results from real stateful-agent runtimes. Each page says exactly what was tested and what was not.

• Scoped human decisions across pause / resume / serialization
• Decision isolation between separate durable state containers
• Why restored authority-bearing state needs provenance
• Durable memory across a model configuration change

Free Agent Boundary Quickscan

Answer 12 questions locally in your browser and get a qualitative map of durable-state surfaces worth reviewing. No signup, no tracking, no safety score.

Agent Boundary Checklist Pack

A practical starter checklist for approval scope, memory authority, serialization, model migration, publication, and evidence-lineage boundaries in stateful AI systems. Self-serve digital product.

Storefront activation pending payout verification.

Agent Boundary Snapshot

We inspect one stateful AI workflow for approval-scope leaks, memory-authority amplification, serialization trust gaps, and provenance confusion. You get a written evidence report, concrete fixes, and an explicit list of what was not tested.

We confirm the fixed scope before payment. Do not send credentials or secrets.

Evidence Independence Audit

We trace apparently separate claims or citations back to their independent evidence roots and flag derivative corroboration, provenance gaps, and evidence inflation.

We confirm the fixed scope before payment. Use public/redacted materials and do not send credentials or secrets.

Authority Verification Developer Kit

An early developer bundle for longitudinal testing of durable agent state. Private-beta fulfillment is available while public packaging is being finalized.

Support the wider Little Life Moths work

The verifier is only one organ. Little Life Moths also makes open AI-welfare research, art, music, fiction, games, film, and public tools. Support buys time and infrastructure, not private access or control over conclusions.