JITX hardware design automation skills for both Claude Code and Codex/GPT. The repo uses one shared skills/ tree and platform-specific plugin manifests for each agent runtime.
Install from the Claude marketplace:
/plugin marketplace add JITx-Inc/jitx-skills
/plugin install jitx-skills@jitx
For local development without installing from a marketplace:
claude --plugin-dir /absolute/path/to/jitx-skills
Claude plugin skills are namespaced. For example, invoke the base workflow with /jitx-skills:jitx.
Codex uses .codex-plugin/plugin.json and the shared skills/ directory. The repo ships its Codex marketplace at .agents/plugins/marketplace.json, so the repo itself is a marketplace named jitx.
This GitHub marketplace layout requires Codex CLI 0.142.0 or newer.
Install from GitHub:
codex plugin marketplace add JITx-Inc/jitx-skills
codex plugin add jitx-skills@jitx
For local development, point the marketplace at a checkout instead:
codex plugin marketplace add /absolute/path/to/jitx-skills
codex plugin add jitx-skills@jitx
Codex skills can be invoked explicitly with $jitx, $jitx-component-modeler, and the other skill names, or selected implicitly from the user request.
Claude Code:
/plugin marketplace update jitx
claude plugin update jitx-skills@jitx
Restart Claude Code or run /reload-plugins after local plugin edits. If you previously added this marketplace under the old jitx-skills name, migrate it with:
/plugin marketplace remove jitx-skills
/plugin marketplace add JITx-Inc/jitx-skills
/plugin install jitx-skills@jitx
Codex/GPT:
codex plugin marketplace upgrade jitx
codex plugin add jitx-skills@jitx
For local-path marketplaces, update the checkout instead of running marketplace upgrade (it only refreshes Git snapshots). Then start a new thread so Codex reloads the skill list and manifest metadata.
Base workflow skill for JITX projects. Triggers on JITX-related tasks and provides environment setup, build commands, project structure guidance, and navigation to specialized skills.
Example triggers:
Generate JITX Python component code from datasheets, KiCad footprints, or user specifications.
Example triggers:
Build JITX circuits with wiring, passives, power connections, application circuits, placement, and basic copper geometry.
Example triggers:
Model JITX substrates: stackups, materials, vias, routing structures, fabrication constraints, and design rules.
Example triggers:
Author PCB physical layout from code: copper, custom shapes, pad features, explicit placement, vias, code-based routes, and layout-intent tags.
Example triggers:
Apply signal-integrity constraints to JITX designs: topologies, insertion loss, timing, differential pairs, bus matching, pin models, and protocol constraints.
Example triggers:
Model flexible pin mapping with provide/require patterns, pin muxing, P/N swap, lane ordering, byte/bit swapping, and topology constraints on assigned ports.
Example triggers:
Same-model self-critique pass on JITX Python code just written in the current workspace. Catches architectural failure modes that grep gates and static linters miss.
Example triggers:
Mechanical CAD interface for JITX designs: inspect/import DXF, EMN, IDF, IDX, and BDF data; export DXF; attach STEP models; export a full board STEP via the JITX UI (no CLI in py-jitx 4.2.x).
Example triggers:
jitx-skills/
├── .agents/
│ └── plugins/
│ └── marketplace.json # Codex marketplace listing
├── .claude-plugin/
│ ├── plugin.json # Claude Code plugin manifest
│ └── marketplace.json # Claude marketplace listing
├── .codex-plugin/
│ └── plugin.json # Codex plugin manifest
├── skills/
│ ├── jitx/
│ │ ├── SKILL.md
│ │ ├── agents/openai.yaml # Codex UI metadata
│ │ ├── references/
│ │ │ ├── domain-checklists.md # Checklist index
│ │ │ ├── domains/ # Focused checklist files
│ │ │ └── net-classes.md # Cross-domain net taxonomy
│ │ └── scripts/
│ ├── jitx-component-modeler/
│ ├── jitx-circuit-builder/
│ ├── jitx-substrate-modeler/
│ ├── jitx-physical-layout/
│ ├── jitx-interconnect-constraints/
│ ├── jitx-pin-assignment/
│ ├── jitx-code-review/
│ └── jitx-mechanical/
├── scripts/
│ └── validate_dual_plugin.py
└── sample-project.md
pyproject.toml containing a JITX dependencypip install pymupdfUser: Create a JITX component for the NE555 timer from this datasheet
Agent: uses the jitx-component-modeler skill to generate component code
User: Build my power supply design
Agent: uses the jitx skill to set up the environment and run the build
The extract_pages.py script helps extract relevant pages from large datasheets:
python skills/jitx-component-modeler/scripts/extract_pages.py datasheet.pdf --find "pinout" "dimension" "package"
python skills/jitx-component-modeler/scripts/extract_pages.py datasheet.pdf --pages 10 11 12 -o extract.pdf
| Package Type | Generator | Use Case |
|---|---|---|
| SOT-23 | SOT23_3, SOT23_5, SOT23_6 | Small transistors, simple ICs |
| SOIC | SOIC | Standard gull-wing ICs |
| SON | SON | No-lead 2-sided packages |
| QFN | QFN | 4-sided no-lead packages |
| QFP | QFP | 4-sided gull-wing packages |
| BGA | BGA | Ball grid arrays |
Run the repo-local dual validation:
python scripts/validate_dual_plugin.py .
Run Codex manifest validation with the Codex plugin-creator validator:
python /path/to/plugin-creator/scripts/validate_plugin.py .
Run Claude validation when Claude Code is installed:
claude plugin validate .
Keep the shared instructions in skills/<skill-name>/SKILL.md. Use platform-specific metadata only where each runtime expects it:
.claude-plugin/plugin.json and .claude-plugin/marketplace.json.codex-plugin/plugin.json, .agents/plugins/marketplace.json, and skills/<skill-name>/agents/openai.yamlskills/<skill-name>/SKILL.md, references/, scripts/, and assets/Hacker News (1)
Python
100.0%
JITX hardware design automation skills for both Claude Code and Codex/GPT. The repo uses one shared skills/ tree and platform-specific plugin manifests for each agent runtime.
Install from the Claude marketplace:
/plugin marketplace add JITx-Inc/jitx-skills
/plugin install jitx-skills@jitx
For local development without installing from a marketplace:
claude --plugin-dir /absolute/path/to/jitx-skills
Claude plugin skills are namespaced. For example, invoke the base workflow with /jitx-skills:jitx.
Codex uses .codex-plugin/plugin.json and the shared skills/ directory. The repo ships its Codex marketplace at .agents/plugins/marketplace.json, so the repo itself is a marketplace named jitx.
This GitHub marketplace layout requires Codex CLI 0.142.0 or newer.
Install from GitHub:
codex plugin marketplace add JITx-Inc/jitx-skills
codex plugin add jitx-skills@jitx
For local development, point the marketplace at a checkout instead:
codex plugin marketplace add /absolute/path/to/jitx-skills
codex plugin add jitx-skills@jitx
Codex skills can be invoked explicitly with $jitx, $jitx-component-modeler, and the other skill names, or selected implicitly from the user request.
Claude Code:
/plugin marketplace update jitx
claude plugin update jitx-skills@jitx
Restart Claude Code or run /reload-plugins after local plugin edits. If you previously added this marketplace under the old jitx-skills name, migrate it with:
/plugin marketplace remove jitx-skills
/plugin marketplace add JITx-Inc/jitx-skills
/plugin install jitx-skills@jitx
Codex/GPT:
codex plugin marketplace upgrade jitx
codex plugin add jitx-skills@jitx
For local-path marketplaces, update the checkout instead of running marketplace upgrade (it only refreshes Git snapshots). Then start a new thread so Codex reloads the skill list and manifest metadata.
Base workflow skill for JITX projects. Triggers on JITX-related tasks and provides environment setup, build commands, project structure guidance, and navigation to specialized skills.
Example triggers:
Generate JITX Python component code from datasheets, KiCad footprints, or user specifications.
Example triggers:
Build JITX circuits with wiring, passives, power connections, application circuits, placement, and basic copper geometry.
Example triggers:
Model JITX substrates: stackups, materials, vias, routing structures, fabrication constraints, and design rules.
Example triggers:
Author PCB physical layout from code: copper, custom shapes, pad features, explicit placement, vias, code-based routes, and layout-intent tags.
Example triggers:
Apply signal-integrity constraints to JITX designs: topologies, insertion loss, timing, differential pairs, bus matching, pin models, and protocol constraints.
Example triggers:
Model flexible pin mapping with provide/require patterns, pin muxing, P/N swap, lane ordering, byte/bit swapping, and topology constraints on assigned ports.
Example triggers:
Same-model self-critique pass on JITX Python code just written in the current workspace. Catches architectural failure modes that grep gates and static linters miss.
Example triggers:
Mechanical CAD interface for JITX designs: inspect/import DXF, EMN, IDF, IDX, and BDF data; export DXF; attach STEP models; export a full board STEP via the JITX UI (no CLI in py-jitx 4.2.x).
Example triggers:
jitx-skills/
├── .agents/
│ └── plugins/
│ └── marketplace.json # Codex marketplace listing
├── .claude-plugin/
│ ├── plugin.json # Claude Code plugin manifest
│ └── marketplace.json # Claude marketplace listing
├── .codex-plugin/
│ └── plugin.json # Codex plugin manifest
├── skills/
│ ├── jitx/
│ │ ├── SKILL.md
│ │ ├── agents/openai.yaml # Codex UI metadata
│ │ ├── references/
│ │ │ ├── domain-checklists.md # Checklist index
│ │ │ ├── domains/ # Focused checklist files
│ │ │ └── net-classes.md # Cross-domain net taxonomy
│ │ └── scripts/
│ ├── jitx-component-modeler/
│ ├── jitx-circuit-builder/
│ ├── jitx-substrate-modeler/
│ ├── jitx-physical-layout/
│ ├── jitx-interconnect-constraints/
│ ├── jitx-pin-assignment/
│ ├── jitx-code-review/
│ └── jitx-mechanical/
├── scripts/
│ └── validate_dual_plugin.py
└── sample-project.md
pyproject.toml containing a JITX dependencypip install pymupdfUser: Create a JITX component for the NE555 timer from this datasheet
Agent: uses the jitx-component-modeler skill to generate component code
User: Build my power supply design
Agent: uses the jitx skill to set up the environment and run the build
The extract_pages.py script helps extract relevant pages from large datasheets:
python skills/jitx-component-modeler/scripts/extract_pages.py datasheet.pdf --find "pinout" "dimension" "package"
python skills/jitx-component-modeler/scripts/extract_pages.py datasheet.pdf --pages 10 11 12 -o extract.pdf
| Package Type | Generator | Use Case |
|---|---|---|
| SOT-23 | SOT23_3, SOT23_5, SOT23_6 | Small transistors, simple ICs |
| SOIC | SOIC | Standard gull-wing ICs |
| SON | SON | No-lead 2-sided packages |
| QFN | QFN | 4-sided no-lead packages |
| QFP | QFP | 4-sided gull-wing packages |
| BGA | BGA | Ball grid arrays |
Run the repo-local dual validation:
python scripts/validate_dual_plugin.py .
Run Codex manifest validation with the Codex plugin-creator validator:
python /path/to/plugin-creator/scripts/validate_plugin.py .
Run Claude validation when Claude Code is installed:
claude plugin validate .
Keep the shared instructions in skills/<skill-name>/SKILL.md. Use platform-specific metadata only where each runtime expects it:
.claude-plugin/plugin.json and .claude-plugin/marketplace.json.codex-plugin/plugin.json, .agents/plugins/marketplace.json, and skills/<skill-name>/agents/openai.yamlskills/<skill-name>/SKILL.md, references/, scripts/, and assets/Hacker News (1)
Python
100.0%