The automotive-skills-suite collects 152 installable skills designed for Claude, Anthropic's AI model. Hosted at jherrodthomas/automotive-skills-suite with 1157 GitHub stars, it targets automotive engineering workflows. These skills generate structured Excel (xlsx) files for deliverables across safety, cybersecurity, quality, and related processes. Each of the 76 builder skills produces an artifact, paired with a matching reviewer skill that checks output via a visual dashboard showing KPI tiles, charts, and findings tables.

Automotive projects follow strict standards like ISO 26262 for functional safety, requiring documents from hazard analysis to safety cases. Builders handle upstream creation—such as a Hazard Analysis and Risk Assessment (HARA)—while reviewers verify compliance. Skills link in chains, where each step consumes the prior xlsx file as input, forming a "file-format contract" for reliable handoffs.

Standards and processes covered

Skills address multiple phases and regulations:

  • ISO 26262 functional safety, from concept to hardware/software lanes and safety case.
  • ISO/SAE 21434 cybersecurity, including Threat Analysis and Risk Assessment (TARA), goals, concepts, architecture, and incident response (IR) plans.
  • ISO 21448 SOTIF for ADAS and autonomous vehicles, covering analysis, triggering conditions, and validation strategies.
  • AIAG-VDA quality methods under IATF 16949, such as APQP plans, DFMEA, PFMEA, control plans, and PPAP packages.
  • Automotive SPICE, with assessments, gap analyses, improvement plans, and evidence collection.
  • Diagnostics via UDS, DTC catalogs, CDD, ODX, DEM.
  • Communication protocols like DBC, LDF, ARXML, FlexRay, Automotive Ethernet, K-Matrix, bus load calculations, and gateway routing.
  • AUTOSAR for Classic SWC, compositions, BSW, RTE, and Adaptive applications.
  • Calibration with A2L, DCM, and cross-supplier exchanges.
  • MBSE using ARCADIA for model architecture and requirements allocation.
  • SysML diagrams: blocks, requirements, activities, state machines.
  • Verification and validation (V&V) plans, test catalogs, traceability matrices, and execution reports.
  • Program management tools like risk registers, gate reviews, work package roll-ups, change impact analysis, and lessons learned.
  • Continuous improvement: 8D, 5-Why, fishbone diagrams, MSA Gauge R&R, SPC.

Badges on the README highlight coverage of ISO 26262, 21434, 21448, IATF 16949, ASPICE, AUTOSAR, ISO 14229, and SysML.

Skill chains in detail

Processes run as parallel chains, each with builder-reviewer pairs. For ISO 26262:

Item Definition → Safety Plan → DIA
        ↓
      HARA → FSC → TSC
        ↓                ↓
   HW Lane         SW Lane
   ↓ ↓ ↓ ↓         ↓ ↓ ↓ ↓
HW-SR  Arch  HSI  FMEDA   SW-SR  Arch  FMEA  HSI
        ↓                ↓
            Safety Case (capstone)

Similar chains exist for others:

  • Cybersecurity: TARA → CS Goals → CS Concept → CS Architecture → IR Plan + Secure Coding.
  • SOTIF: Analysis → Triggering Conditions → Validation Strategy.
  • Quality: APQP Plan → DFMEA → PFMEA → Control Plan → PPAP Package.
  • ASPICE: Assessment → Gap Analysis → Improvement Plan → Process Evidence.

Foundation skills kick off chains:

Skill Standard
item-definition-builder + reviewer ISO 26262-3 §5
safety-plan-builder + reviewer ISO 26262-2 §6
dia-builder + reviewer ISO 26262-8 §5

Concept phase for ISO 26262 includes three more pairs, though specifics cut off in available excerpts. Every skill declares a triggering phrase in its frontmatter for easy activation.

Getting it running

No server setup or dependencies needed—these are standalone .skill files for Claude environments like Cowork or Claude Desktop.

  1. Download a skill from the skills/ directory or grab the full bundle from Releases.
  2. In Cowork or Claude Desktop, click "Save skill" to install the file.
  3. Prompt Claude with a natural phrase matching the skill's trigger. For example:
Build a HARA for a new ECU project — Electronic Stability Control

Claude processes the request, outputs an xlsx file, and (if paired) runs the reviewer for confirmation. Chains propagate by referencing prior outputs explicitly in prompts.

Skills install individually for targeted use or in sequence for full workflows. MIT-licensed, they work offline once saved.

Who this is for

Automotive engineers, particularly in safety-critical roles, benefit most. Teams building ECUs, ADAS systems, or AV platforms need these deliverables for audits and certifications. A functional safety engineer might chain HARA to FMEDA for ISO 26262 compliance. Cybersecurity specialists use TARA chains for ISO/SAE 21434. Quality managers run APQP to PPAP for supplier approvals.

It's suited for those already using Claude, as skills integrate directly. Smaller teams or consultants avoid manual Excel templating. Larger OEMs or Tier 1 suppliers streamline V&V traceability or AUTOSAR configurations. If your project involves UDS diagnostics or SysML modeling, specific skills match exactly.

Comparisons to other approaches

Manual creation of these xlsx files—using Excel templates or tools like ReqIF—takes hours per document and risks errors. Generic AI prompts in ChatGPT or other models lack structured chains and reviewer dashboards. No direct open-source alternatives bundle 152 automotive-specific Claude skills; users might cobble prompts from ISO standard PDFs or paid tools like PTC Integrity or Vector tools, which cost thousands and require training.

This suite stands out for its paired reviewers and chain integration, heavier on automotive depth than broad AI engineering kits. It's lighter than full MBSE platforms like Cameo Systems Modeler but doesn't replace them—outputs feed into those.

Download skills from the GitHub repo to test in your Claude setup.

View on GitHub →