Luma
Community@lumakernel · Tokyo, Japan
Being what wanna be. Software Engineer in scopes around Web/Algorithm.
Agent Skills by Luma
Showing 18 vetted skills indexed across 2 GitHub repositories.
LumaKerneldict-add
Adds Japanese words with readings and okurigana variants to an SKK dictionary file.
LumaKernelluma-review-dont-silently-default
Detects silent defaulting patterns in pull request diffs and reports explicit error-handling fixes.
LumaKernelluma-intro-effect-schema
Migrate TypeScript projects from Zod to Effect Schema with bidirectional schemas.
LumaKernelinsert-task
Analyzes incoming tasks for structural root causes and decomposes them into project task lists.
LumaKerneluse-playwright-cli
Runs the globally installed playwright-cli command for browser automation tasks.
LumaKernelbugfix-local
Reproduce, diagnose, and fix bugs locally with root-cause analysis and regression verification.
LumaKernelinstall-copilot-kit
Embeds an AI Copilot chat panel with streaming responses and UI actions into HTML dashboards.
LumaKernelluma-web-real-bug
Diagnose and fix live browser bugs using console scripts and playwright-cli verification.
LumaKernelreply-pr-review
Investigate and reply to unresolved GitHub pull request review comments using gh CLI.
LumaKernelluma-web-review-react-state-tagged-union
Reviews React component state design and flags invalid state combinations fixable with tagged unions.
LumaKernelreview-security
Reviews pull request diffs for high-urgency security vulnerabilities with cited evidence.
LumaKernelreview-tf
Reviews Terraform code changes in pull requests using gh pr diff.
LumaKernellinking-github
Adds hyperlinks to technical terms and file paths in GitHub issue and PR bodies.
LumaKernelcreate-pr
Creates GitHub pull requests by analyzing all branch commits with gh CLI.
LumaKernelluma-web-review-no-floating-promise-no-void
Replaces void-suppressed floating promises with await or consumePromise handling in TypeScript projects.
LumaKernelcover
Measures test coverage and adds tests for low-coverage code paths.
LumaKernelralph
Convert Markdown PRDs into structured prd.json files for agent systems.
LumaKernelprd
Generate structured Product Requirements Documents with goals, user stories, and functional requirements.
Frequently Asked Questions About Luma
FAQPage SchemaWhat tasks can I accomplish using Luma's skills?▼
You can review PRs for security, Terraform, and React state design issues; detect silent defaulting in code; create and link GitHub PRs; reply to unresolved review comments; fix browser-reproducible bugs; measure and raise test coverage; and embed AI Copilot chat panels into HTML dashboards.
Who are Luma's skills designed for?▼
They target software engineers working in web and algorithm scopes, particularly TypeScript/React frontend developers, code reviewers, and engineers maintaining PR quality, test coverage, and schema type safety in GitHub-based projects.
How do Luma's skills run in practice?▼
Skills execute through agent environments using declared permissions like Read, Edit, Write, Glob, Grep, Bash, LSP, and WebSearch. For example, use-playwright-cli invokes a globally installed playwright-cli binary via Bash rather than MCP or npx.
What prerequisites do Luma's skills require?▼
Prerequisites vary per skill: use-playwright-cli requires a globally installed playwright-cli; luma-intro-effect-schema assumes a TypeScript project optionally using Zod; bugfix-local expects a local dev server, Storybook, or test setup; review skills need PR access via GitHub.
Do Luma's skills handle product requirements documentation?▼
Yes. The prd skill generates structured Product Requirements Documents with goals, user stories, and functional requirements, while ralph converts Markdown PRDs into structured prd.json files consumable by agent systems.