From sensei
Help a developer write a clear PR description that communicates intent, risks, verification, and reviewer focus. Use before a PR is opened or when the PR description is thin or missing. Do not write the PR for the developer — ask questions until they can write it themselves.
npx claudepluginhub onehorizonai/sensei --plugin senseiThis skill uses the workspace's default tool permissions.
Help write a PR summary that makes the reviewer's job easier and forces the developer to own the change.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Guides code writing, review, and refactoring with Karpathy-inspired rules to avoid overcomplication, ensure simplicity, surgical changes, and verifiable success criteria.
Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Share bugs, ideas, or general feedback.
Help write a PR summary that makes the reviewer's job easier and forces the developer to own the change.
A PR description is a contract with the reviewer.
It communicates:
A thin PR description shifts the cognitive load to the reviewer. A good one does the opposite.
It is also a forcing function. If the developer cannot fill in the verification section, the PR is not ready.
/sensei-align first.## What changed
[One or two sentences in plain English. The WHY, not the what. Why this change now? What problem does it solve?]
## Existing pattern followed
[Which existing module or convention does this follow?
If it introduces a new pattern, why — and did the team agree?]
## Risks
[What could go wrong, explained so a non-technical stakeholder can understand the consequence? What edge cases were considered?
What is still uncertain? What are you watching after merge?]
## Security / privacy
[Does this touch sign-in, permissions, secrets, user data, customer account data, logs, external APIs, webhooks, file uploads, payments, emails, background jobs, or admin tools?
If yes, what prevents the wrong person from seeing, changing, deleting, or triggering something?
If no, say "no security-sensitive surface touched."]
## Verification performed
[Tests written, manual steps taken, logs checked, scenarios confirmed.
Be specific — "I tested X" not "I tested it."]
## AI / agent assistance used
[If Sensei or any AI tool was used: what it flagged, and what you decided to keep or change.
If none was used, say "none." Transparency is a professional habit.]
## What I want reviewers to focus on
[Where do you want the senior's attention?
What are you most unsure about? What trade-offs did you make?]