crit
Skill15 starsChanged 4 months agoMITView on GitHub
Request a Crit — a narrated design review of the running app from the user. Use when you want human design/product feedback on UI you built or changed, when the user should look at the app and react to it, or when the user says things like "crit", "get my feedback", "review this with me", or "I'll talk you through it". Opens the app in a local review runner; returns a timestamped transcript of spoken feedback with pointer and source-file context.
- Reads credentials
- Installs packages
3 other projects use a version of this.
Adopted by
Most recently taken up by
- lorem-dev/duologtheir own version, first seen Nov 15, 2026
- vmarafetti/crittheir own version, first seen Oct 17, 2026
- tomasz-tomczyk/skillstheir own version, first seen May 17, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- vmarafetti/crit★ 9same name
- tomasz-tomczyk/skills★ 7same name
- lorem-dev/duolog★ 4same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 54
If you run it on a workspace root with several apps, it exits with MONOREPO_APP_AMBIGUOUS and lists the candidate app dirs — pick the one …
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Crit — ask the user for a narrated app review
Crit opens the user's browser on a mirrored copy of the current app with a recording widget. The user narrates feedback out loud while pointing with their mouse. You get back a merged, timestamped review transcript where references like "this button" are resolved to concrete elements with source file, line, and component — enough context to act without guessing.
1. Ensure the CLI is available
command -v crit || npm install -g design-crit(If a global install isn't possible, npx -y design-crit review ... works too.)
2. Check for API keys
Crit needs OPENAI_API_KEY (speech-to-text; also covers the merge). ANTHROPIC_API_KEY is
optional (Claude then does the merge). Keys are read from your shell environment first, then
from the project's .env/.env.local. If neither has a key, tell the user and either ask them
to add one or run with --mock-ai (interaction context only, placeholder transcript).
3. Request the Crit
Tell the user: "I'm requesting a Crit — your browser will open. Press Start, talk through your feedback while pointing with your mouse, and press Stop when you're done."
Then run (long timeout — the command blocks while the user reviews, often several minutes):
crit review --source . --jsonTo land the reviewer on a specific page — do this whenever the feedback you want concerns a
particular route (e.g. you just changed the checkout flow) — pass --path:
crit review --source . --json --path /checkoutMonorepos: point --source at the app package, not the workspace root — crit detects the
enclosing workspace automatically, mirrors the whole repo, and runs the dev server from the app:
crit review --source apps/web --json
# equivalently: crit review --source . --app-dir apps/web --jsonIf you run it on a workspace root with several apps, it exits with
MONOREPO_APP_AMBIGUOUS and lists the candidate app dirs — pick the one the user's feedback
concerns and rerun with --app-dir. If the app depends on sibling services (an API server),
make sure those are already running; crit only starts the app's own dev server.
4. Consume the result
stdout is a single JSON object:
"status": "completed"— readfinal_transcript(timestamped Markdown review). Theartifacts.review_jsonfile has the same content structured:issues[](title, evidence, timestamp, url),timeline[],suggested_followups[]. Interaction context includes source anchors likesrc/components/Hero.tsx:42 <Hero>— use them to open the exact code the user pointed at."status": "cancelled"— the user declined; do not relaunch without asking."status": "error"— reporterror.codeanderror.message; partial artifacts (raw speech transcript, interaction timeline) may still be listed inartifacts.
5. Act on it
Summarize the issues back to the user, map each to its source location (the anchors are in the transcript), propose concrete changes, and implement the unambiguous ones. Where the review says the target was unclear, ask the user rather than guessing. After making changes, offer another Crit to verify.
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
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.

