Public research, written for responsible release.
Pentoma publishes sanitized research briefs and report samples that explain evidence standards, remediation patterns, and safety boundaries without exposing raw artifacts or exploit-ready detail.
Research library
Approved public records only.
The catalog is intentionally small: one sanitized report sample plus approved briefs for web authorization, source review, and AI agent red teaming.
Sanitized Pentoma assessment report
A public-safe report excerpt showing how Pentoma frames scope, confirmed evidence, remediation, and compliance mapping without exposing customer artifacts.
- Ecosystem
- Application security reporting
- Status
- sample
Safe public validation for web authorization research
A brief on publishing authorization research with enough detail to be credible while keeping tenant, role, and replay mechanics private.
- Ecosystem
- Web applications and APIs
- Status
- approved
Publishing source review findings without leaking code
Guidance for turning source-code security review into useful public research while keeping proprietary lines, paths, and secrets out of view.
- Ecosystem
- Source-code review
- Status
- approved
Public evidence for AI agent red teaming
A publication model for AI agent findings that explains observed behavior, tool boundaries, and mitigations without sharing jailbreak recipes.
- Ecosystem
- LLM and agentic systems
- Status
- approved
Publication guardrails
Credible does not need to mean dangerous.
Pentoma public research should make the validation model clear while preserving customer confidentiality, disclosure boundaries, and safe remediation focus.