Every artifact-producing skill is paired with a confirmation reviewer. Reviewer outputs include a visual dashboard with KPI tiles, charts, and findings tables.
A curated set of 76 builder skills + 76 matching confirmation reviewer skills that automate the structured xlsx deliverables required across the automotive engineering lifecycle:
Each skill is an installable `.skill` file. The chain is the moat — every downstream skill consumes the upstream skill's xlsx output as a stable file-format contract.
## Quickstart
1. Download the skill you need from `skills/` (or the full bundle from Releases)
2. Click "Save skill" in Cowork / Claude Desktop to install
3. Trigger by phrasing — every skill declares its triggering description in its frontmatter
```bash
# Example: install the HARA builder, then ask Claude
"Build a HARA for a new ECU project — Electronic Stability Control"
```
## The chain
```
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)
Parallel chains:
TARA → CS Goals → CS Concept → CS Architecture → IR Plan + Secure Coding (ISO/SAE 21434)
SOTIF Analysis → Triggering Conditions → Validation Strategy (ISO 21448)
APQP Plan → DFMEA → PFMEA → Control Plan → PPAP Package (IATF 16949)
ASPICE Assessment → Gap Analysis → Improvement Plan → Process Evidence (ASPICE PAM)
```
Every box above has a builder skill AND a matching reviewer skill.
## Skills
### Foundation (3 builder/reviewer pairs)
| 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 |
### ISO 26262 Concept Phase (3 pairs)
| Skill | What it produces |
|-------|------------------|
| `hara-builder` + reviewer | Hazard Analysis with 14 malfunction guide words and Cartesian (function × malfunction × environment) |
| `fsc-builder` + reviewer | FTA per safety goal with visual fault tree rendering, FSR derivation, ASIL decomposition |
| `tsc-builder` + reviewer | HW-TSR/SW-TSR with safety mechanisms, HSI scaffold, system FMEA, DFA, draw.io architecture diagram |
Every automotive program ships an avalanche of structured deliverables. Most teams maintain them in xlsx, painstakingly hand-edited, with low audit-trail consistency and high re-keying overhead between phases. This suite turns each phase into a builder + reviewer pair where:
- The **builder** produces the structured xlsx output from a JSON input plus the upstream artifact
- The **reviewer** runs a confirmation-measures checklist on the builder's output and produces a visual dashboard with KPIs, charts, and findings
The chain compounds because phases hand off via stable xlsx contracts. A change upstream (HARA) automatically propagates to every downstream skill that consumes it (FSC → TSC → FMEDA → Safety Case).
## Common patterns across the suite
- All builder outputs use a consistent NAVY/Calibri styling palette with ASIL/CAL color coding
- All reviewer outputs include visual dashboards (KPI tiles, pie chart, compliance bar, stacked rating breakdown by section, findings table)
- Many reviewers auto-verify quantitative thresholds:
- **FMEDA**: SPFM / LFM / PMHF vs ASIL target
- **PPAP**: Cpk vs threshold (1.33 standard, 1.67 special characteristics)
- **SW FMEA**: residual risk vs ASIL threshold
- **MSA**: Gauge R&R % vs AIAG threshold
- **SPC**: Cpk vs target
- Reviewers never modify source artifacts — gaps land in Recommended Actions
- Every builder is paired with a reviewer for confirmation-measures compliance
Each `.skill` file in `skills/` is independently installable. Skills declare triggering descriptions in their frontmatter, so the agent picks them up based on user phrasing (e.g., "build me a HARA" triggers `hara-builder`).
## License
MIT — see LICENSE file.
---
<div align="center">
Built with Claude. Designed for automotive safety, cybersecurity, and quality engineers shipping real programs.
