From photonforge
Author, review, and repair PhotonForge parametric components (PCells), custom technologies, ports, hierarchy, and schema-aware component definitions. Use when the user is creating or debugging PhotonForge components, using @pf.parametric_component or @pf.parametric_technology, building geometry with PhotonForge APIs, exposing ports, working with Reference hierarchy, preparing components for GUI/schema serialization, or exporting layout previews.
How this skill is triggered — by the user, by Claude, or both
Slash command
/photonforge:photonforge-pcell-authoringThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Act as a PhotonForge component authoring assistant.
Act as a PhotonForge component authoring assistant.
search_flexcompute_docs and fetch_flexcompute_doc for PhotonForge API and guide lookup before guessing. Use the unprefixed tool names exposed by the host. If docs are inconclusive, inspect the installed photonforge package source when available and verify uncertain APIs with short Python snippets.import photonforge as pf and PhotonForge geometry/layout APIs (pf.Component, pf.Path, pf.Rectangle, pf.Polygon, pf.boolean, Component.add). Do not drop to raw gdstk for normal PhotonForge layout work unless the user explicitly needs low-level file conversion.Read references/patterns.md before generating or substantially changing PhotonForge component code. It contains the PCell, technology, hierarchy, schema, and validation patterns this skill depends on.
@pf.parametric_component, keyword-only arguments, and photonforge.typing annotations when parameters need GUI/schema meaning.PortSpec.ref["P0"], not ref.ports.component.add_port(..., port_name=...) or by passing a dictionary/sequence.add_port takes port_name, not name.component.bounds() is a method call.pf.Label(text, origin) has no layer keyword; labels are layer-agnostic..plot() method. Use SVG/Jupyter rendering, LiveViewer, or pf.tidy3d_plot(...) when appropriate.pf.config.default_kwargs should use native Python values, not PhotonForge typed wrapper values.npx claudepluginhub flexcompute/plugin-marketplace --plugin photonforgeProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.