Kicad Review logo

Kicad Review

Community
mixelpixx
kicad-review

Design review and validation workflow for KiCAD projects via MCP tools. Triggers on: "review my design", "check for errors", "audit", "DRC", "ERC", "find problems", "design review", "is this ready", "validate", "check my schematic", "check my PCB", "what's wrong", "run checks", "pre-fab review".

Overview

Publishermixelpixx
RepositoryKonnect
Skill namekicad-review
Stars
944
Forks
127
Bundled files
2
LicenseAGPL-3.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by mixelpixx on GitHub. Read the source before you install it.

Installation

Install the Kicad Review AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/mixelpixx/Konnect/tree/main/crates/konnect/assets/skills/kicad-review
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/mixelpixx/Konnect.git /tmp/Konnect
mkdir -p .claude/skills
cp -r /tmp/Konnect/crates/konnect/assets/skills/kicad-review .claude/skills/kicad-review
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Kicad Review in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Kicad Review on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Kicad Review is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

KiCAD Design Review & Validation Workflow

This skill guides Claude through systematic design review of a KiCAD project using MCP tools. ALL checks are performed through MCP tools — never parse .kicad_sch or .kicad_pcb files directly.


Toolset Loading

Load the required toolsets for design review:

load_toolset('sch_analysis')     # find_orphan_items, find_shorted_nets, find_single_pin_nets
load_toolset('verification')     # run_drc; check_clearance anchor/courtyard spacing; rules
load_toolset('sch_export')       # run_erc
load_toolset('pcb_export')       # get_drc_violations
load_toolset('manufacturing')    # validate_for_manufacturing
load_toolset('design_review')    # audit_decoupling, audit_connections, audit_power_rails, etc.

Optional (for deeper analysis):

load_toolset('sch_analysis')     # get net info, trace connections, inspect components
load_toolset('pcb_routing')      # query_traces, get_nets_list

Always call get_active_toolsets() first to see what is already loaded.

Evidence hierarchy

Judge findings in this order:

  1. Exact design requirements and manufacturer datasheets.
  2. Direct KiCad ERC/DRC and saved or exported connectivity.
  3. Direct Konnect net, short, pad, trace, via, unrouted, and inventory evidence.
  4. Aggregate review and manufacturing summaries.
  5. Heuristic orphan, single-pin, decoupling, protection, and best-practice findings.

A weaker finding may ask a question; it does not override stronger contradictory evidence. Any required check that did not run, returned impossible coverage, or remains inconsistent with stronger evidence makes the verdict INCOMPLETE.

DRC violation items carry owner and ownership_status. That is direct evidence, not a heuristic: owner.kind: "footprint" means the offending geometry belongs to owner.reference's own artwork, so the remedy is a footprint or rule change rather than a placement change — the finding itself stands either way. When ownership_status is not "resolved", ownership is unknown and owner is null; corroborate with list_board_footprint_graphics rather than assuming the board owns it.

References by review branch

  • Read references/design-checklist.md for a comprehensive or pre-fabrication review. Mark an item only from evidence collected in this run.
  • Read references/error-taxonomy.md when classifying a finding or assigning the final verdict. Direct ERC, DRC, and connectivity evidence outrank heuristic classifications.

Quick Checks (Escalating Severity)

Run these first — they are fast and catch the most critical issues.

Level 1: Structural Integrity

find_orphan_items()

Finds floating wires, labels, and symbols not connected to anything. Treat the result as a heuristic candidate list and corroborate it with direct connectivity or ERC before calling an item a defect.

Level 2: Critical Net Issues

find_shorted_nets()

Detects nets that are connected together but should not be. A shorted net means:

  • Two different net labels on the same wire
  • Power rails bridged unintentionally
  • Signal nets merged by accident

A confirmed unintended short is critical. Resolve disagreement with requirements or direct ERC/connectivity evidence before assigning severity.

Level 3: Suspicious Connections

find_single_pin_nets()

A one-pin net is a heuristic review candidate:

  • Incomplete wiring (forgot to connect the other end)
  • Orphan net labels (typo in name, so it does not match)
  • Leftover stubs from deleted components

Formal Checks

ERC — Electrical Rules Check

run_erc()

Checks schematic-level rules:

  • Pin type conflicts (output driving output, unconnected inputs)
  • Power pin connections
  • Missing no-connect flags
  • Duplicate reference designators
  • Missing net connections

Review each violation. Some can be waived (e.g., intentional unconnected pins marked with no-connect flag).

DRC — Design Rules Check

get_drc_violations()

Checks PCB-level rules:

  • Clearance violations (copper-to-copper, copper-to-edge)
  • Minimum trace width violations
  • Minimum drill size violations
  • Unrouted connections (incomplete routing)
  • Zone fill issues
  • Courtyard overlaps

Every DRC error must be resolved or explicitly justified before manufacturing.


Design Audits

These go beyond rule checking — they evaluate design quality and best practices.

The standalone schematic audits and check_bom_health default to the supplied file only. When the supplied file is a hierarchy root, pass schematic_scope: "hierarchy" to cover every reachable sheet instance. Read status, coverage, and diagnostics before interpreting a hierarchy result; missing or cyclic child references make the result incomplete. Reused child files have one result per KiCad sheet instance, identified by the sheet_instance_path response field.

Decoupling Audit

audit_decoupling(schematic, schematic_scope="hierarchy")

Checks:

  • Every IC power pin has a bypass capacitor
  • Capacitor is placed close to the pin (PCB proximity)
  • Appropriate capacitor values (100nF ceramic minimum)
  • Bulk capacitance present for high-current ICs

Connection Audit

audit_connections(schematic, schematic_scope="hierarchy")

Checks:

  • All expected connections are made
  • No nets with unexpected fan-out
  • Signal integrity basics (termination on long traces)
  • Pull-up/pull-down resistors where required (I2C, reset pins, enable pins)

Power Rail Audit

audit_power_rails(schematic, schematic_scope="hierarchy")

Checks:

  • All power rails have proper source (regulator, connector, etc.)
  • Current capacity matches expected load
  • Voltage levels are consistent (no 3.3V device on 5V rail)
  • Power sequencing considered for multi-rail designs
  • Power flags present (avoids ERC false positives)

Manufacturing Audit

audit_manufacturing()

Checks:

  • All footprints are fab-house compatible
  • Pad sizes meet minimum requirements
  • Silkscreen readability
  • Test point accessibility
  • Fiducial marks present (for SMT assembly)
  • Mechanical clearances around mounting holes

Full Review Shortcut

run_design_review()

Runs the aggregate design audits and produces a consolidated report. Use it to organize findings, not as a substitute for the direct ERC, DRC, and connectivity checks above.

Read status, coverage, and diagnostics before interpreting the findings. If status is partial or failed, the verdict is INCOMPLETE — review could not evaluate the full design. Report that verdict verbatim, explain the diagnostics and unevaluated coverage, and do not describe the design as ready, passing, clean, or looking good. Findings gathered before the coverage gap are still valid and should still be reported.

Collect direct run_erc, get_drc_violations, short, and connectivity evidence separately. Then compare the aggregate findings with that stronger evidence and report any disagreement.


Severity Classification

CRITICAL — Must fix before manufacturing

FindingWhy Critical
Shorted netsShort circuit on the board, may damage components
Missing ground connectionCircuit will not function
Reversed polarity on power ICImmediate destruction on power-up
Unrouted netsMissing connections on fabricated board
DRC clearance violationMay cause electrical short on fab board
Power pin unconnectedIC will not operate
Wrong voltage on IC power pinExceeds absolute maximum, destroys part

WARNING — Should fix, design risk

FindingWhy a Warning
Missing decoupling capacitorNoise susceptibility, possible oscillation
No test points on key signalsCannot debug in production
No ESD protection on connectorsVulnerable to ESD damage in the field
Single-point-of-failure netsNo redundancy for critical signals
Pull-up/pull-down missingFloating input, unpredictable behavior
Tight clearances (near DRC limit)Higher fab defect rate

SUGGESTION — Improvement opportunities

FindingWhy a Suggestion
Consolidate passive valuesFewer unique BOM lines, lower assembly cost
Add net labels to unnamed netsImproves schematic readability
Missing silkscreen designatorsHarder to assemble and debug manually
Components could be closerShorter traces, better signal integrity
Consider bulk capacitor additionBetter transient response on power rails
Add board revision markingTraceability for manufacturing runs

Reporting Format

Present findings grouped by severity with actionable fix suggestions:

## Design Review Results

### CRITICAL (X issues) — Must fix

1. **[Finding title]**
   - Location: [component reference or net name]
   - Issue: [what is wrong]
   - Fix: [specific action to take using MCP tools]

### WARNING (X issues) — Should fix

1. **[Finding title]**
   - Location: [component reference or net name]
   - Issue: [what is wrong]
   - Fix: [specific action to take]

### SUGGESTION (X items) — Optional improvements

1. **[Finding title]**
   - Detail: [what could be better]
   - Action: [suggested improvement]

### Summary
- Critical: X (must resolve)
- Warnings: X (recommended)
- Suggestions: X (optional)
- Verdict: [LOOKS GOOD / NEEDS ATTENTION / NOT READY / INCOMPLETE]
- Coverage status: [complete / partial / failed]
- Coverage diagnostics: [none, or each unevaluated sheet/object/audit]

Review Workflow

Quick Review (5-minute check)

  1. find_shorted_nets() — catch fatal issues
  2. run_erc() — schematic rule check
  3. get_drc_violations() — PCB rule check
  4. Report findings

Full Review (comprehensive)

  1. Load all review toolsets
  2. Run direct short/connectivity checks, run_erc, and get_drc_violations
  3. run_design_review() — aggregate audit suite
  4. Check status, coverage, and diagnostics; never approve an incomplete review
  5. Reconcile aggregate or heuristic findings with stronger direct evidence
  6. Classify all gathered findings by severity
  7. Present report with fix suggestions
  8. Offer to fix CRITICAL issues immediately

Pre-Manufacturing Review

  1. Full review (above)
  2. Run validate_for_manufacturing(), then inspect verdict, issues, and drc against the handler's limited contract; it does not replace outline, drill, silkscreen, artifact, BOM/CPL, or order-preview acceptance
  3. Verify BOM completeness
  4. Check part availability (if targeting specific fab house)
  5. Final verdict: ready to manufacture or not

Rules

  1. Never skip quick checks — find_shorted_nets catches the worst bugs fast
  2. Classify every finding — severity helps the user prioritize
  3. Provide specific fixes — name the MCP tool and parameters to resolve each issue
  4. Run DRC after fixes — verify that corrections did not introduce new violations
  5. Do not approve a design with CRITICAL issues — even if the user says "it's fine"
  6. Load toolsets first — check get_active_toolsets() and load what you need
  7. Save before reviewing — ensures checks run against current state
  8. Offer to fix — after reporting, offer to use MCP tools to resolve issues
  9. Re-run after fixes — always verify fixes resolved the issue and created no new ones
  10. Document waivers — if user explicitly waives a warning, note it in the report
  11. Never soften INCOMPLETE — partial or failed coverage is not a passing review

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Kicad Review AI skill do?

Design review and validation workflow for KiCAD projects via MCP tools. Triggers on: "review my design", "check for errors", "audit", "DRC", "ERC", "find problems", "design review", "is this ready", "validate", "check my schematic", "check my PCB", "what's wrong", "run checks", "pre-fab review".

Why use Kicad Review on TypingMind?

Because you install it once and use it with any model. Kicad Review is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Kicad Review in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/mixelpixx/Konnect/tree/main/crates/konnect/assets/skills/kicad-review. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Kicad Review?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Kicad Review?

As many as you like. As long as a model supports skills, you can use Kicad Review with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Kicad Review AI skill free?

Yes. It is published on GitHub by mixelpixx under the AGPL-3.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇