datasheet-reader

The datasheet-reader skill is a thin, focused helper that uses the `pcb scan` command-line tool to convert datasheets or technical PDFs into readable markdown. It solves the practical problem of working with dense electronics datasheets: rather than parsing a raw PDF, the skill runs a scan that materializes both a cached PDF and a markdown rendering, then directs the agent to work from the markdown artifact where text, tables, and image links are far easier to consume.

The workflow is minimal by design. You run `pcb scan` against a local `.pdf` path or an `http(s)` URL, capture the path reported on the `Markdown:` output line, read that markdown file, and follow linked images only when a task depends on figures, diagrams, or tables. The skill advises preferring URLs when they succeed, notes that the cached PDF is not a package artifact and should not be copied into the repository by default, and recommends the minimal invocation without optional flags unless a task explicitly requires them. It also gives simple failure guidance: report the failure briefly and fall back.

This skill targets hardware and PCB engineers using the diodeinc `pcb` toolchain who frequently need to pull component specifications out of manufacturer datasheets or extract values from KiCad `Datasheet` properties. It is a read-only convenience wrapper — it fetches and renders reference documents rather than modifying designs — so it is well suited to research-heavy component-selection and reference tasks within a PCB workflow.