GitHub Repository

Functional safety toolkit for AI-assisted development — ISO 26262, IEC 61508, IEC 60730, Automotive SPICE, MISRA-C:2012

4 starsPython

Torasan – Open-source functional safety automation framework

Name: Torasan – Open-source functional safety automation framework
Availability: InStock
Author: enzo_ferrari

by enzo_ferrari·Mar 8, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidNiche GemBig BrainSolve My Problem

41 Claude skills + V-model process for ISO 26262 automation—niche but genuinely useful.

Strengths

•Encodes real domain expertise (FMEA, safety concept generation, SPICE evidence) into reusable prompts instead of SaaS gatekeeping.
•Complete working example (BLDC motor controller) with generated artifacts proves the framework produces actual compliance output.
•Targets genuine pain: one-person safety teams doing manual traceability matrices—solves a scaling problem in supply-chain engineering.

Weaknesses

•Claude Code dependency limits portability; unclear if skills transfer cleanly to other LLM APIs despite 'LLM-agnostic' claim.
•No live demo or generated document samples in repo—must clone and run locally to evaluate output quality.

Post Description

I built TORASAN (Toolkit for Organized Risk Analysis and Safety Assurance Navigation) to automate the tedious parts of functional safety development.

If you've ever worked with ISO 26262, IEC 61508, or Automotive SPICE, you know how much time goes into process compliance — traceability matrices, FMEA, safety concepts, SRS documents, test design, static analysis, etc. Most of it is structured enough to be automated, yet engineers still do it manually.

TORASAN is a process framework + 40 CLI skills that handle: - V-model phase execution with gate verification - FMEA generation with severity/occurrence/detection ratings - Safety concept creation (safety goals → technical requirements) - Software requirements specification generation - Traceability matrix management - MISRA C:2012 compliance checking (cppcheck + clang-tidy + flawfinder) - Automotive SPICE process assessment (16 processes, targeting Level 2)

It's built on Claude Code but designed to be LLM-agnostic — the process definitions, templates, and knowledge base work independently of the AI backend.

Supports ISO 26262:2018, IEC 61508, IEC 60730, and MISRA C:2012.

Two repos available: - TORASAN-dev: full git history (157 commits showing the development journey) - TORASAN-public: clean sanitized version for forking

GPLv3 licensed. Feedback welcome.