)]}'
{
  "log": [
    {
      "commit": "289525bcecca82247e0908c5ff1f213213d9c9a4",
      "tree": "ffdc00894b39af359cba7f481c05e9a5b7b405f7",
      "parents": [
        "828161b106396bf87a4c481895781f1d1d732e0a"
      ],
      "author": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Tue Jul 08 04:03:02 2025 +0000"
      },
      "committer": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Mon Jul 07 21:13:27 2025 -0700"
      },
      "message": "sketch/termui: handle PortMessageType messages more\n\n- Added case for loop.PortMessageType in termui message handling switch\n- Port messages now display with 🔌 emoji and clean formatting\n- Follows same pattern as other message types (ErrorMessageType, BudgetMessageType, etc.)\n\n- Added loop.PortMessageType to AddMultipleUnion call in go2ts.go\n- Regenerated TypeScript types to include \u0027port\u0027 in CodingAgentMessageType union\n- Webui components now have proper type safety for port messages\n\nCo-Authored-By: sketch \u003chello@sketch.dev\u003e\nChange-ID: s7e1f4b705a66a9d5k\n"
    },
    {
      "commit": "112b92376c97b8da64cc1c954896957cfc479f3d",
      "tree": "1497eb8b2927f1334d712aabb19625db11a148bf",
      "parents": [
        "d203b7de7d49cc5da03440d5a00b2efd0f8bb2c8"
      ],
      "author": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Fri May 23 11:26:33 2025 -0700"
      },
      "committer": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Thu May 29 12:50:41 2025 -0700"
      },
      "message": "loop: add todo checklist\n\nThis should improve Sketch\u0027s executive function and user communication."
    },
    {
      "commit": "272a90ee1a74bda5618d4866e03f4b7067947784",
      "tree": "9baf4f84ce80b1c7073a95b6959f6dd11ab3b48b",
      "parents": [
        "d3ac112a45111abf0e57c327d55e2cc66a136abb"
      ],
      "author": {
        "name": "Philip Zeyliger",
        "email": "philip@bold.dev",
        "time": "Fri May 16 14:49:51 2025 -0700"
      },
      "committer": {
        "name": "Philip Zeyliger",
        "email": "philip@bold.dev",
        "time": "Fri May 16 14:51:40 2025 -0700"
      },
      "message": "Add Monaco diff-view, the saga ...\n\nI set out to use Monaco to support the diff view. diff2html is lovely,\nbut there were a ton of usability improvements I wanted to make (line\nnumbers not making things double spaced, choosing which diff, editing\nthe right-hand side), and it seemed a dead end. Furthermore, Phabricator\nand Gerrit\u0027s experience is that diffs should be shown file by file,\nbecause you\u0027ll inevitably see a diff with a file that\u0027s too large, and\nthe GitHub PR view often breaks on big changes... so I wanted to show\nfiles diff-by-diff, with \"infinite\" context when unchanged sections are\nexpanded. So...\n\nUltimately, all of this was sketch-coded over maybe 30 Sketch sessions.\nI threw away a lot of branches. My git reflog is a superfund site.\n\nPrompting whole-hog didn\u0027t work. Or, rather, it made significant\nprogress, but something very serious wouldn\u0027t work, and I couldn\u0027t\nfigure out what, and nor could Sketch.\n\nInstead, I started by adding a new webcomponent that was just a\nplaceholder. Then, using https://rodydavis.com/posts/lit-monaco-editor,\nI nudged Sketch into adding Monaco to it. Sketch pulled out:\n\n   You\u0027re right, I should properly read the blog post before implementing the\n   solution. Let me check the referenced blog post.\n\nI worked heavily in the demo environment at first, but here I ran into\nthe issue that we have two different esbuild systems: one is vite and\none is esbuild.go, and they\u0027re configured differently enough.\n\nMonaco is unusable and confusingly so when its CSS isn\u0027t loaded. The right\nway to load it, I\u0027ve found, is via\n\n  @import url(\u0027./static/monaco/min/vs/editor/editor.main.css\u0027);\n\nI spent more time than I care to admit noticing that originally\nthis wasn\u0027t relative, and when we use a skaband setting, the\npaths need to be relative-aware.\n\nThe paths to the various workers need to be similarly correctly placed.\n\nGetting Sketch to build demo data but not put testing code into production\ncode was tricky. (I threw away a lot of efforts and factories and singletons...)\n\nWhen I set out to do the git commit selection, I wanted to do a bunch of\nbackend /git/* handlers. These were easy enough to code in sketch. I had\nto convince Sketch to put them in git_tools.go and not in the agent.\nIt doesn\u0027t really matter: these functions to parse git are pretty stateless,\nbut it\u0027s less work to have them separate. Sketch was mediocre at writing\ntests for them. Did you know that our container has an older version\nof git that doesn\u0027t have the same options to decorate ref names? Yeah, nor did\nI.\n\nHandling unstaged changes was fun. git diff --raw shows unstaged files\nas having identity 0000. Ideally we\u0027d be using jj and there\u0027d be\na synthetic commit, but instead uncommitted-possible files are read\nby content.\n\nA real big challenge was getting the Monaco view to use the right vertical and\nhorizontal space. I did this many, many times. I don\u0027t claim to understand flex\nand the virtual dom, and :host, and all the interactions. It would fix one\nthing and break another. The chat window would shrink. The terminal would\nshrink.\n\nScreenshot support was excellent. I eventually added paste support just so\nthat I could expedite my workflow, and Sketch coded that easily on the first\npass with minor feedback.\n\nI learned the hard way that Safari\u0027s support for WebComponents/shadow\ndom in its web inspector is rough. See https://fediverse.zachleat.com/@zachleat/114518629612122858\n\nI also learned the hard way that Chrome doesn\u0027t use fonts loaded in CSS\nin a shadow dom. That\u0027s why the codicon font had to be in the global\nstyle sheet.\n\nKudos to John Reese who kindly allowed me, a long time ago, to adapt a\nshell script he had at work to look over diffs into https://github.com/philz/git-vimdiff.\nThat\u0027s the inspiration for having the \"new code\" be editable when you\u0027re\nreviewing it; why shouldn\u0027t it be!?!\n\nThere are a handful of follow up tasks:\n\n* We lose state when we switch to the Chat view and back.\n* Need URL-based support for where we are.\n* Maybe need shortcut keys to move between diffs and changes.\n* Maybe need caching or look-ahead for downloading the next or previous\n  file.\n* We spend too much vertical real estate on all the diff selections;\n  could we scroll it out of the way, collapse it, tighten it, etc.\n* The workers sometimes throw errors into the console. I think they\u0027re\n  harmless and merely need to be caught and suppressed.\n* Needing to commit changes when things are saved is weird. Should we\n  commit automatically? Amend the previous commit? Have a button for\n  that? Show the git dirty state?\n* Our JS bundle is big. We could maybe delay loading the monaco bundle\n  to help.\n\nThanks for coming to my TED talk.\n"
    },
    {
      "commit": "485afc6975b853aa670b5087c323fa3df3df0672",
      "tree": "291bbf09de69cec72c13f652f3d9be81f35451e4",
      "parents": [
        "6234a8d9d229904994f49a5edacca05e54b26d1f"
      ],
      "author": {
        "name": "Sean McCullough",
        "email": "banksean@gmail.com",
        "time": "Mon Apr 28 14:28:39 2025 -0700"
      },
      "committer": {
        "name": "Autoformatter",
        "email": "bot@sketch.dev",
        "time": "Fri May 02 22:12:51 2025 +0000"
      },
      "message": "tool_use: add multiplechoice support to Agent\n\nThis implements the \"dumb\" approach - the tool itself just tells\nthe llm that it rendered the options to the user, and it\u0027s done.\n\nIf the user selects one of the options, we paste its response text\ninto the chat input textarea on the frontend.  The user is of\ncourse free to ignore the question or the options presented.\n\nThis keeps no association between user response and the original\ntool_use block that solicited it from the user. I.e. the user\nresponse message doesn\u0027t include the original tool_use_id value\nit it.  It looks as though the user typed it by hand.\n"
    },
    {
      "commit": "4f84ab729ddbf428b0e891940f08f70b4edee05c",
      "tree": "f2e52e4a01c188ada1f5acf8b2a013029b999495",
      "parents": [
        "44f9b4cec11e269a52fbfc099989ab425b8e125f"
      ],
      "author": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Tue Apr 22 16:40:54 2025 -0700"
      },
      "committer": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Fri May 02 12:57:44 2025 -0700"
      },
      "message": "all: support openai-compatible models\n\nThe support is rather minimal at this point:\nOnly hard-coded models, only -unsafe, only -skabandaddr\u003d\"\".\n\nThe \"shared\" LLM package is strongly Claude-flavored.\n\nWe can fix all of this and more over time, if we are inspired to.\n(Maybe we\u0027ll switch to https://github.com/maruel/genai?)\n\nThe goal for now is to get the rough structure in place.\nI\u0027ve rebased and rebuilt this more times than I care to remember.\n"
    },
    {
      "commit": "bc4947e41d50b5e1a8089600808047715b828305",
      "tree": "beef5beca05e5714a3c1f16822bf1b2f388c2436",
      "parents": [
        "d9d4581952edb59469a8402910783570be852fb8"
      ],
      "author": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Wed Apr 30 18:23:11 2025 -0700"
      },
      "committer": {
        "name": "Josh Bleecher Snyder",
        "email": "josharian@gmail.com",
        "time": "Wed Apr 30 18:23:11 2025 -0700"
      },
      "message": "cmd/go2ts: fix typo\n\nIt is now an ex-🍁.\n"
    },
    {
      "commit": "d9d4581952edb59469a8402910783570be852fb8",
      "tree": "7b06cf74a10d3d812152c21fc72ced395a3e5024",
      "parents": [
        "48c84c9a4df29b1e51c7fdca81053dd9cd0c2446"
      ],
      "author": {
        "name": "Sean McCullough",
        "email": "banksean@gmail.com",
        "time": "Wed Apr 30 16:53:41 2025 -0700"
      },
      "committer": {
        "name": "Sean McCullough",
        "email": "banksean@gmail.com",
        "time": "Wed Apr 30 18:00:55 2025 -0700"
      },
      "message": "webui: add current Agent state to call-status tooltip\n"
    },
    {
      "commit": "2032b1c1971ceb85ca14b20273a3783729fba3e3",
      "tree": "0486e9222643ffcbbd34286148f4a7913a169668",
      "parents": [
        "4f50a68ac73677c0022b2b3da8b4667cee01c11b"
      ],
      "author": {
        "name": "Philip Zeyliger",
        "email": "philip.zeyliger@gmail.com",
        "time": "Wed Apr 23 19:40:42 2025 -0700"
      },
      "committer": {
        "name": "Philip Zeyliger",
        "email": "philip.zeyliger@gmail.com",
        "time": "Wed Apr 23 19:40:42 2025 -0700"
      },
      "message": "Move webui from /loop/webui to /webui\n\nThanks, perl (and git mv):\n\n\tperl -pi -e s,loop/webui,webui,g $(git grep -l loop/webui)\n"
    },
    {
      "commit": "d9f1337ec3317a60df50d8ba4eefb448473d62fa",
      "tree": "67b60b05d8e2e3fb01293f68dfc5d08600a28dda",
      "parents": [
        "db8c5abcc18082bcbf47247e8d50889a32e65e06"
      ],
      "author": {
        "name": "Sean McCullough",
        "email": "banksean@gmail.com",
        "time": "Mon Apr 21 15:08:49 2025 -0700"
      },
      "committer": {
        "name": "Sean McCullough",
        "email": "banksean@gmail.com",
        "time": "Mon Apr 21 15:21:40 2025 -0700"
      },
      "message": "webui: auto-generate types.ts from go structs\n"
    }
  ]
}