net-convert

name	net-convert
description	Convert KiCad netlist (.net) files to structured Markdown and manage schematic knowledge base with vector + keyword hybrid retrieval. Use when: user needs to parse .net netlist files, extract MCU pin assignments, build schematic knowledge base, or query netlist information.

NET-CONVERT - KiCad Netlist Conversion & Knowledge Base Skill

Automatically convert KiCad netlist files (.net) to structured Markdown reports with MCU pin assignment tables, and build a searchable schematic knowledge base.

When to Use

✅ USE this skill when:

• "Convert .net netlist files to markdown"
• "Parse the schematic netlist and show MCU pins"
• "Build knowledge base from netlist files"
• "What pins does the MCU use for PWM?"
• "Show me the pinout for [MCU model]"
• "Query schematic knowledge base"

When NOT to Use

❌ DON'T use this skill when:

• Input is not a KiCad .net file
• Need to convert PDFs/Excel/Pack → use doc-convert skill instead
• Need to generate code → use fae skill instead
• Only need to read a single text file → use read tool directly

Supported Formats

Format	Extension	Converter	Output
KiCad Netlist	.net	converter.py → scripts/netlist_to_md.py	.md (structured report)

Directory Structure

skills/net-convert/
├── SKILL.md                     # This file
├── converter.py                 # Batch netlist conversion tool
├── kb_manager.py                # Schematic knowledge base manager
├── requirements.txt             # Python dependencies (shared with doc-convert)
├── scripts/
│   └── netlist_to_md.py        # KiCad .net → Markdown parser
├── retriever/                   # Retrieval engine (imports from doc-convert)
│   └── __init__.py
└── kb/
    └── config.json              # Knowledge base configuration

Paths

# Netlist source search
SEARCH_ROOT="/Users/sinomcu/.openclaw/workspace/fae_input/sources/"

# Converted output
OUTPUT_DIR="/Users/sinomcu/.openclaw/workspace/fae_input/cache/schematics/"

# Knowledge base indexes (shared with doc-convert)
KB_ROOT="/Users/sinomcu/.openclaw/workspace/fae_input/"

Usage Examples

Example 1: Convert a Single Netlist

User: "Convert this netlist file to markdown"

Assistant: (uses net-convert skill)

python3 /Users/sinomcu/.openclaw/workspace/skills/net-convert/converter.py \
  -i /path/to/MC8059_12VPUMP-G-20260323_V10.net \
  -o /Users/sinomcu/.openclaw/workspace/fae_input/cache/schematics/

Output: cache/schematics/MC8059_12VPUMP-G-20260323_V10.md

• MCU pin assignment table
• BOM list
• Signal net details
• Structured pin data (KB-optimized)

---

Example 2: Batch Convert All Netlists

User: "Convert all netlist files in the project"

Assistant: (uses net-convert skill)

python3 /Users/sinomcu/.openclaw/workspace/skills/net-convert/converter.py \
  -i /Users/sinomcu/.openclaw/workspace/fae_input/sources/ \
  -o /Users/sinomcu/.openclaw/workspace/fae_input/cache/schematics/ \
  -r

Output: All .net files converted to structured Markdown

---

Example 3: Build Schematic Knowledge Base

User: "Build knowledge base from the converted netlists"

Assistant: (uses net-convert skill)

python3 /Users/sinomcu/.openclaw/workspace/skills/net-convert/kb_manager.py build

Output: Vector + keyword indexes updated with netlist pin data

---

Example 4: Query Schematic Knowledge Base

User: "What pins does MC8059 use for PWM?"

Assistant: (uses net-convert skill)

python3 /Users/sinomcu/.openclaw/workspace/skills/net-convert/kb_manager.py query "MC8059 PWM pin assignment"

Output: Relevant pin assignment entries from the schematic knowledge base

---

Example 5: Check Status

User: "Show schematic knowledge base status"

Assistant: (uses net-convert skill)

python3 /Users/sinomcu/.openclaw/workspace/skills/net-convert/kb_manager.py status

Commands Reference

Conversion

# Convert single file
python3 converter.py -i design.net -o cache/schematics/

# Batch convert (recursive)
python3 converter.py -i fae_input/sources/ -o fae_input/cache/schematics/ -r

# Force re-conversion (ignore cache)
python3 converter.py -i design.net -o cache/schematics/ -f

Knowledge Base

# Build index
python3 kb_manager.py build

# Query
python3 kb_manager.py query "your question"

# Status
python3 kb_manager.py status

# Clear
python3 kb_manager.py clear

Output Structure (Markdown Report)

Each converted netlist produces a structured report with:

1. Design Information — source file, tool version, date
2. MCU Pin Assignment Table — grouped by function:
   • 电源/参考电压 (Power/Reference)
   • 总线/通信接口 (Bus/Communication)
   • PWM / 驱动输出 (PWM/Drive Output)
   • 模拟输入/采样 (Analog Input/Sampling)
   • 定时器/计数器 (Timers/Counters)
   • GPIO (可能上拉/限流)
   • 悬空 (Unconnected)
3. 器件清单 (BOM) — grouped by type
4. 信号网络详细连接 — all signal nets with component connections
5. 电源网络连接详情 — power net details
6. MCU 引脚结构化数据表 — KB-optimized, one pin per line

Inference Confidence Levels

Pin function inference uses confidence markers:

Badge	Level	Meaning
✅	High	Standard power/ground names (GND, VDD, 5V, etc.)
⚠️	Medium	Known signal patterns (SPI, UART, PWM, ADC, etc.)
❓	Low	Topology-based inference (RC network, single resistor, etc.)
—	None	No pattern matched, raw net name preserved

Integration with Other Skills

FAE Skill

FAE → net-convert (parse netlist) → schematics/ → FAE (read pinout) → generate code

Doc-Convert (Shared KB)

net-convert → convert netlists → cache/schematics/ → kb_manager.py build → shared indexes/
                                                              ↑
doc-convert → convert PDFs → cache/text/ ─────────────────────┘

Both skills write to the same fae_input/indexes/ knowledge base, enabling unified RAG retrieval across datasheets, requirements, and schematic pin assignments.

Dependencies

Uses the same dependencies as doc-convert:

cd /Users/sinomcu/.openclaw/workspace/skills/doc-convert
pip3 install -r requirements.txt

Key dependencies:

• chromadb - Vector database
• whoosh - Keyword search engine
• sentence-transformers - Embedding models

No additional dependencies required for net-convert (the netlist parser is pure Python).

Known Limitations

• KiCad format only — Currently supports KiCad Eeschema .net format. Altium, Eagle, or other formats are not supported.
• Single MCU focus — Designed for single-MCU boards. Multi-MCU designs will only extract the primary MCU.
• No schematic rendering — This is a text parser, not a graphical viewer.

---

This skill parses KiCad netlist files and provides structured pin assignment data for FAE code generation and RAG retrieval.