OSIIX Library
OSIIX Library catalogs research reports, books and learning material, practical guides, and case studies and engineering notes. It organizes knowledge from actual investigation, verification, development, and operations so you can discover it by material type.
Research Report · V2 Full Report / Brief
A research report that frames ransomware not simply as encryption malware, but as an organizational risk combining intrusion, credential abuse, control of administrative privileges, lateral movement, data theft, recovery disruption, business interruption, and extortion. The Full Report and Brief are available in both English and Japanese.
Book Overview
An English overview of Backup / Recovery / Archive Fundamentals. English Chapter 1 and Chapter 2 remain available.
SSRF Series · Chapter 1 · English
Chapter 1 explains the basic structure of SSRF, server-side communication, reachability, privileges and trust, and the difference between input and the actual communication destination.
SSRF Series · Chapter 2 · English
Chapter 2 explains why SSRF defenses are difficult through the gap between a defensive decision and the communication ultimately performed, organized as a six-perspective analytical model.
Version 1 · English / 日本語
A learning resource that organizes the AI fundamentals most relevant to AI-assisted development, including models, generation, observability, RAG, and verification.
Reference · 48 terms
A reference guide to 48 terms concerning AI reliability, errors, evaluation, agent behavior and failure, and human or security-related controls.
Public Draft v0.15
A public guide to methods mars70s actually tried in AI-assisted development and operational support, and found useful for reducing drift and rework. It covers change boundaries, secret material, production operations, verification, records, and stopping procedures with practical examples.
Operational proposal
A practical framework for resuming work without mixing up projects, Git, local worktrees, or approval states by keeping the target, evidence, execution path, and approval boundaries distinct.
Technical guide / implementation note
A Safety Harness that checks the Repository Root, Write Set, Stage Set, Branch, push destination, and Secret Scan before execution, and stops when the observed state differs from expectations.
Case study
Separates what public sources confirm, assess, and leave unknown, including the “less than 10 hours” and “around two weeks” figures, evidence paths for AI use, and the parallelization hypothesis.
Case study
A troubleshooting record that separates website-side issues from AI-side fetch controls
Case study
An examination of how AI can satisfy an instruction yet still reach an outcome the human did not want, and a practical proposal to compare candidate success states before planning.
Development record
A record of trial and error building a security monitoring platform with AI assistance: contract-driven architecture, the read-only boundary, Producer/Consumer separation, and how UNKNOWN is handled.
Development record
A development record that reframes CRLF/LF raw hash mismatches as a missing Agent interpretation contract rather than a Repository normalization problem — including how an initial hypothesis was disproved by a Git history investigation, and the minimal contract adopted instead.
Development record
A development record examining, through a real integration case, where to place the change boundary when incorporating an existing web tool into another site. It also covers the idea of validating application health and public-route health separately.
Case study / technical record
A technical record of a body-size limit that initially depended on Content-Length, the move to enforcing actual received bytes at the ASGI layer, and why source regression testing and runtime validation were treated as separate checks.
Case study
A qualitative case study drawing on development and operations records that remain traceable to their source repositories and commit history — examining two independent evaluation axes, Scope Boundary and Verification Boundary, through a bounded recovery outcome, a case where a proposal's scope expanded, and a record that shows where the Verification Boundary should be drawn.
Operational case study
A read-only review of 29 AGENTS.md files across 22 repositories, showing why line count and shared text alone did not justify changing any of them.
Case study
A practitioner case study of two Windows Codex CLI execution failures, separating process-creation denial from resource-authorization denial through evidence-preserving investigation.
Case study
A case study contrasting two website-operation cases by stage, reversibility, blast radius, and failure cost, while preserving the limits of the available evidence.
Case study
A Git-history-based case study of an incomplete SSOT migration, a legacy mutable-state copy that remained inconsistent with the canonical state, and a structural fix that moved §4 state ownership back to WORK_STATUS references.
This guide presents methods mars70s tried while using AI in a personal development environment. It is not a general standard or the only correct approach, but one example to adapt to your own environment. AI assistance was used for wording and layout.