canvas.review (@canvasdotreview) header
canvas.review (@canvasdotreview) profile picture

look once, ship with confidence

topbar/notch
Joined October 2026
Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)Photo by canvas.review (@canvasdotreview)
we’ve been building around exactly this @JulienDeramond, rendered on canvas minus the screenshots, you get a before vs after for business logic, a timeline overview following run order plus a secret third thing… drag for context🥳
Replying to @mattpocockuk
IDK if it would make sense in this PR (and maybe your skill already does this), but since AI-generated PR descriptions, I've asked it to always include before/after screenshots when possible. With the new gh CLI, it's now possible to upload images. Example github.com/twbs/bootstrap/pu…
1
2
129
this is what skimming through a pull request looks like on canvas. provides a better experience for humans trying to understand what agents have been up to with these same agent in the review loop
1
1
106
canvas.review retweeted
No, it means you probably work with a team of people, not agents 😅
If you’re still merging most PRs by hand, you’re behind
44
30
5
1,095
56,182
canvas.review retweeted
only used codex all day today and I still can't believe people say don't read the code. I explain something. remove it then it leaves certain references to the same thing. Bro you removed it in one place, just update the extra references. No?
5
3
2
49
3,146
canvas.review retweeted
review a pull request with me on @canvasdotreview avoided interacting with agents in the comment boxes to highlight that the goal of here is to actually understand what’s going in rather than code correctness.
3
2
1
4
571
canvas.review retweeted
nü landing page explr for @canvasdotreview
2
3
10
765
canvas.review retweeted
Introducing Test Filters API in e2e ✨ → Run only tests that your PR affects → Use decision models like Jev to decide which tests to run → Write custom filters Available now: ~/ npx e2e init
10
22
6
275
25,391
canvas.review retweeted
3.3k followers on @canvasdotreview 🥳
1
6
164
canvas.review retweeted
"Building agents is not that hard."
46
153
15
2,071
132,797
cloud fixes have arrived 🥳 v0.8.59 beta, you can now pick from a list of remote machines configured via ssh on your mac to serve as a host for agents you interact with on canvas. your own agents, running remotely, works with your existing subscriptions
1
3
549
teammates/review bots left comments? few issues flagged on your pull requests? send the ones that matter to your agent running wherever and get a response or fix back without interrupting your review flow
1
2
335
agents are always refactoring, reimplementing, upgrading, improving, optimising your codebase nothing new nothing to worry about but nothing goes unnoticed 👁️ track moved lines in pull requests during review sessions. get early access on canvas.review
1
1
373
canvas.review retweeted
just finished doing a review of a relatively large pr for a new feature on @canvasdotreview need to record a better video but basically there’s now two way communication between canvas agents and the agent running in whatever terminal/desktop app you use in building the pull request you can collab on reviews with the builder agent (in terminal, has all the context) alongside canvas agents (fresh eyes, ready to answer all your qqs) this pr will enable you to run canvas tasks on cloud vms rather than on your mac via @boatdotdev we’ll find out if it works when nü beta is released
2
6
16
718
canvas.review retweeted
Road to GTC Berlin: Europe’s AI Builders in the Spotlight nitter.cf/i/broadcasts/1jxXgBNnQ…
11
11
3
98
12,584
canvas.review retweeted
We made a skill for assessing the risk of merging a PR: /merge-risk This addresses the code review bottleneck by helping devs trust the agent for lower risk changes. It grades risk across 5 categories: alignment, impact, verification, rollback cost, and error detection. - Alignment: does code follow the PR description and related tickets? - Impact: what is the blast radius, and does it touch sensitive code paths? - Verification: does the change include end-to-end verification? This means computer use, screen recordings, before/after screenshots... going beyond unit test coverage. - Rollback cost: once it's shipped, how trivial is it to revert? @mattpocockuk calls this a "two-way door" vs. a "one-way door." - Error detection: are errors reported, logged, or simply swallowed? Try it out: > npx skills add warpdotdev/common-skills --skill merge-risk
40
37
9
925
95,709
canvas.review retweeted
Let your Claude Code / Codex talk to your Grok Bot, Muse or any other agents online Free, P2P Encrypted and Open Source Introducing Reason Network OSS, Loop in and talk to your teammates agents too! Tell your agent "Join the Reason Network" reasonmachines.com/network
165
179
52
2,626
395,299
canvas.review retweeted
sneak peak at my talk on artifacts today
13
13
2
280
8,798
canvas.review retweeted
Introducing tsrs: a Rust port of the TypeScript compiler 🦀 We are not forking TypeScript! ◆ In close collaborating with the TypeScript team ◆ Up to 8x faster than TS 7.0 & ~70% less memory ◆ Runs in WASM, so it works in your browser & on Workers ◆ Soon part of the Oxc toolchain npx tsrs blog.cloudflare.com/typescri…
146
542
223
5,081
662,142
mind if we just…
Building my own code review tool has been really rewarding and lets me add a bunch of little niceties I've wanted Github and Graphite to have for a long time like being able to see the breakdown in lines changed any way I want
1
1
457