feat(website): four FPGA tri commands recorded (x7-fasm, fpga-txhold/rxcap, fpga-wire, fpga-jtag)

[gHashTag]

Four new recording pages under apps/website/public/term/, made with tri cast publish (scrub, check, card, X-sized GIF), for the FPGA-engineer series on @t27_dev:

id	command(s)	what it shows
tri-x7-fasm	tri x7-fasm, tri x7-writer	FASM to frames 22/22 byte-identical to fasm2frames (ours 6.31 s this run; fasm2frames 150.7 s at corpus build); frames to .bit 16/16 byte-identical to xc7frames2bit, wrong part refused
tri-fpga-rgmii	tri fpga-txhold, tri fpga-rxcap	the two recorded model runs (2026-09-27/28) printed: TX hold 3445 ps min, RX mid-nibble 10900 ps min, PASS. SDF model, not a board measurement
tri-fpga-wire	tri fpga-usb, tri fpga-wire 115200 921600 3000000	CP2102N direct, no hub; wire rates 115,385 / 923,077 / 3,000,000
tri-fpga-jtag	tri fpga-jtag --decode ..., --self-test	IDCODE decode (XC7A200T, XC7A100T, even code rejected), 17/17 self-test

Every cast: exit codes all 0, home directory scrubbed to ~, 0 key hits. Gallery public/term/index.html rebuilt (22 recordings).

Second commit: check:tools-coverage required the seven trios cards these recordings run to carry their CAST; regenerated with the card generator at the pinned BrowserOS commit, not edited by hand.

{
  "version": 1,
  "head_sha": "4dfd25329f89a675ab1b93a308cc5881d9a48651",
  "summary": "Four FPGA tri commands are recorded for the FPGA-engineer series: a Xilinx 7-series bitstream assembled without Vivado, RGMII margins from nextpnr's SDF, the baud rate actually on the wire, and which chip answered the JTAG cable; seven trios tool cards now name their recording.",
  "changes": [
    "apps/website/public/term/{tri-x7-fasm,tri-fpga-rgmii,tri-fpga-wire,tri-fpga-jtag}/: four new recording pages (index.html, session.cast, session.gif, card.png, meta.json) written by tri cast publish, and the gallery index rebuilt to 22 recordings",
    "apps/website/public/t27/files/specs/tools/trios/tri/{x7-fasm,x7-writer,fpga-txhold,fpga-rxcap,fpga-usb,fpga-wire,fpga-jtag}.t27: CAST added by scripts/tools-from-trios-tri.mjs at the pinned BrowserOS 7366096248df, no hand edits",
    "apps/website/public/tools/spec-tools.json and spec-tools-trios.json: regenerated by agents-from-specs.mjs, 25 cards with a recorded run",
    "apps/website/scripts/agents-from-specs.test.mjs: withCast pinned at 25 instead of 18, with the 17 trios cards from 14 recordings explained"
  ],
  "tests": [
    {
      "command": "npm run check:tools-coverage",
      "status": "passed",
      "result": "every unambiguous recorded command has its CAST; before the regeneration it reported 7 problems, one per card",
      "evidence": "local run of this head, 2026-10-03"
    },
    {
      "command": "npm run check:tools",
      "status": "passed",
      "result": "607 cards typecheck, 25 with a recorded run",
      "evidence": "local run of this head, 2026-10-03"
    },
    {
      "command": "npm run test:tri-commands",
      "status": "passed",
      "result": "exit 0, no failing test",
      "evidence": "local run of this head, 2026-10-03"
    },
    {
      "command": "node --test --test-name-pattern='the committed recordings' scripts/agents-from-specs.test.mjs",
      "status": "passed",
      "result": "1 of 1 pass; the full file before the edit: 52 of 53, the one failure 25 !== 18 on withCast",
      "evidence": "local run of this head, 2026-10-03"
    },
    {
      "command": "npm run check:tools-site && npm run check:agents && npm run check:onboarding && npm run check:subpath-urls && npm run check:shared-core",
      "status": "passed",
      "result": "all five exit 0",
      "evidence": "local run of this head, 2026-10-03"
    },
    {
      "command": "tri cast check on the four new session.cast files",
      "status": "passed",
      "result": "4 of 4: every exit code 0, key hits 0, home directory shown as ~",
      "evidence": "local run before tri cast publish, 2026-10-03"
    },
    {
      "command": "npx vite build (apps/website)",
      "status": "not_run",
      "result": "not run locally; the website CI build covers it",
      "evidence": "no local build was run for this head"
    }
  ],
  "limitations": [
    "tri fpga-txhold and tri fpga-rxcap print the stored logs of two earlier model runs (2026-09-27 and 09-28); the page says it is an SDF model, not a board measurement.",
    "tri fpga-jtag is recorded with --decode and --self-test only: offline decoding, no cable was driven in this recording.",
    "Not recorded: tri x7-audit and tri x7-mutate exit non-zero here because they call a t27c binary inside a worktree that no longer exists.",
    "Nothing was posted from this PR. The X posts are scheduled by hand after each page answers 200 on t27.ai."
  ],
  "tags": [
    "Website",
    "FPGA",
    "Tooling"
  ],
  "blog": {
    "title": "Four FPGA commands, recorded: bitstream, RGMII, baud rate, JTAG",
    "summary": "Four tri commands an FPGA engineer uses daily are now recorded with real output, from assembling a 7-series bitstream without Vivado to naming the chip on the JTAG cable.",
    "outline": [
      "tri x7-fasm turns FASM into frames byte-identical to fasm2frames on 22 of 22 designs, and tri x7-writer turns frames into a .bit byte-identical to xc7frames2bit on 16 of 16.",
      "tri fpga-txhold and tri fpga-rxcap print the RGMII margins of two recorded model runs, tri fpga-wire shows the CP2102N puts 115,385 baud on the wire for a requested 115,200, and tri fpga-jtag names an IDCODE.",
      "Each recording keeps every byte of output; the prompt and typing are staged, and two commands that failed for a missing binary were left out rather than shown."
    ],
    "cast": "term/tri-x7-fasm/session.cast",
    "reproduce": [
      "tri x7-fasm",
      "tri fpga-wire 115200 921600 3000000",
      "tri fpga-jtag --decode 0x03636093"
    ]
  }
}

🤖 Generated with Claude Code