agentleFS
someboxsomebox/pcb-kit/skills/routing/SKILL.md

routing

Skill4 starsChanged 28 days agoMITView on GitHub

Re-route a KiCad board's signals with the negotiated router, `pcbkit reroute`: small parts moved where that helps, every signal net routed together, test pads placed after the routing, repairs, and the old routing and the new one compared. Use when a board with its power copper and ground drawn needs its signals routed or routed again, when an autorouted or hand-routed board has detours, tangles or too many vias, or when comparing two routings of a board. Not for laying power copper, and not for routing a net at a time from a script (pcbkit.route).

18 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use parallelization, prompt-chaining and 1 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • hajekimSkill

    parallelization

    This skill should be used when the user wants to "run tasks in parallel", "concurrent LLM calls", "fan-out and fan-in", "parallel agents", "simultaneous execution", "reduce latency with concurrency", "parallel sub-tasks", "parallel processing", "concurrent agent execution", "run agents simultaneously", "speed up with parallelism", "async agent execution", "batch LLM calls", "parallel workflow", "parallelize agent tasks", or needs to execute multiple independent tasks simultaneously. Also responds to Korean: "병렬 작업 실행", "동시 LLM 호출", "팬아웃 팬인 패턴", "병렬 에이전트", "동시 실행", "동시성으로 지연 시간 감소", "병렬로 실행", "동시에 여러 작업", "병렬 처리", "동시에 돌려줘", "빠르게 병렬로", "여러 작업 동시 실행해줘". Also responds to Japanese: "並列処理", "同時実行", "並列エージェント", "並行LLM呼び出し", "ファンアウト・ファンイン", "複数タスクを同時に実行したい", "処理速度を上げたい", "非同期エージェント", "バッチ処理", "並列で動かしたい" Also responds to Chinese: "并行处理", "并发执行", "并行智能体", "同时调用多个LLM", "扇出扇入模式", "多任务并行运行", "加快处理速度", "异步智能体", "批量LLM调用", "同时运行多个任务".. Apply this skill to design or implement the Parallelization agentic design pattern.

    Also in 3 projects that use this

  • hajekimSkill

    prompt-chaining

    This skill should be used when the user wants to "chain prompts", "build a pipeline", "break down complex tasks into steps", "sequential LLM calls", "multi-step reasoning", "pipeline pattern", "sequential agent pipeline", "multi-step prompt pipeline", "LLM chain", "step-by-step agent", "prompt pipeline", "decompose task into prompts", "stage-by-stage processing", "chained LLM calls", or when a task is too complex for a single prompt and needs to be decomposed into sequential sub-tasks. Also responds to Korean: "프롬프트 체이닝", "체인 방식 에이전트", "단계별 LLM 호출", "복잡한 작업 분해", "순차적 프롬프트 파이프라인", "단계적 작업 처리", "단계별 처리", "순차 처리 에이전트", "파이프라인 만들어줘", "단계적으로 처리해줘", "연속 프롬프트", "체이닝 해줘". Also responds to Japanese: "プロンプトチェーニング", "ステップごとに処理", "順次LLM呼び出し", "パイプライン構築", "複雑なタスクを分解", "段階的エージェント", "連続プロンプト", "LLMパイプライン", "順番に処理してほしい", "チェーン方式エージェント" Also responds to Chinese: "提示词链", "顺序调用LLM", "构建流水线", "分解复杂任务", "逐步处理", "链式提示", "多步骤流水线", "按顺序处理任务", "提示词管道", "帮我一步一步处理".. Apply this skill to design, implement, or explain the Prompt Chaining agentic design pattern.

    Also in 3 projects that use this

  • NVIDIASkill

    testing

    Run Python and Rust tests for physicsnemo-curator using uv + pytest with coverage reporting via pytest-cov, and cargo-nextest for Rust tests. Includes benchmark workflows with pytest-benchmark and criterion.

    Also in 3 projects that use this

The file

Files
SKILL.md

Routing a board

pcbkit reroute takes a placed board and lays its signals again, whoever routed it first, and leaves the result beside the project's board for you to compare: it never writes the board. It is experimental: six boards routed and judged by DRC and by eye, none built yet, so say so when you present a result. docs/ROUTER.md has the method, every setting and the limits.

Steps

  1. Check that it fits. The router lays F.Cu and B.Cu only, and needs the power copper and ground already on the board. A board with signals on inner layers keeps those nets as they are. A board with no copper yet needs its power laid first, by the project's own script or by hand.
  2. Set up once: pcbkit setup-router (needs uv) installs the router's own Python under pcbkit info's tools dir.
  3. Write the settings in the project, e.g. scripts/reroute-<board>.json. Start small: name the board, and add each differential pair to keep.
    {"board": "../boards/driver/driver.kicad_pcb", "keep": ["GND", "USB_DP", "USB_DN"]}
  4. Stop after the job first: pcbkit reroute scripts/reroute-driver.json --until=job. Read what it prints: what was kept, how many nets it will route, and any !! pour warning. A pour that floods the signals' room goes in lift, with its net in keep. Then run it without --until: a two-layer board takes minutes, a four-layer one about an hour.
  5. Read the result: build/reroute/boards/<name>/summary.json has failed connections, DRC of both routings, and the measures of each under lint. Compare lint.old with lint.new, not the new numbers alone.
  6. Look at it. Numbers don't show a bad route: pcbkit netview on the regions round the main IC, or pcbkit reroute page for old and new side by side. Ask whether you would redo it.
  7. Add placement when the routing is close: "place": {"movable": ["R*", "C*"], "fixed": [...], "limits": {...}}, with every decoupling capacitor in fixed or limits (the search knows nets, not loop areas), and test pads that may move in free.
  8. Adopting a result is the user's call, on a branch, with KiCad closed: final.kicad_pcb replaces the board, then pcbkit drc --refill --save, pcbkit check, and a note in LAYOUT.md with the date, the settings and the numbers. The settings file stays in the project, so the routing can be run again.

Reading the output

  • Failed connections in nroute.log's last lines, by net. The worker count acts like a seed, so try another ("route": {"workers": 4} in the settings, and keep it there: the same count repeats the same routes). A net that fails on every run usually has a pocket no via fits in: keep it as drawn.
  • repair.py's lines in the run's output say what it joined after routing. no spot means a pad that needs a hand fix. a piece of N mm2 joined by a via on a power plane means one via now carries that piece: check the current.
  • DRC counts come without silkscreen kinds. via_diameter and drill_out_of_range mean a class wants a bigger via than the job's one size.
  • The measures: vias, mm, back_mm (track on B.Cu), corners90, jogs, tees (joins away from pads), necks (narrow runs), via_in_ic (vias in pin areas), hug_mm (track beside other nets' pads), under_mm (under ICs), detour_mm, crossbacks.

Gotchas

  • A differential pair routed as two nets stalls the negotiation: always keep it.
  • A kept pour above 20 % of the board walls the signals in: the flow warns, and lift and keep are the fix.
  • A two-layer board with a solid ground pour on the back gets more back-layer track than a person would lay: watch back_mm, and look at what happened to the pour.
  • A change to the router itself is judged on the corpus, not on one board: see "Changing the router" in docs/ROUTER.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.