berlysia
Community@berlysia · Tokyo, Japan
Agent Skills by berlysia
Showing 19 vetted skills indexed across 1 GitHub repositories.
berlysiasession-memo
Summarize active sessions into Markdown memos under .tmp/session-memos.
berlysiaoptimize-claude-md
Analyze CLAUDE.md files and generate improvement reports with implementation plans.
berlysiaupdate-docs
Update project documentation to reflect recent code changes.
berlysiadocs-audit
Audit project documentation for code drift and report findings by severity.
berlysiaclarify
Delegate specification elicitation to a sub-agent and summarize final requirements.
berlysiatask-enrich
Enrich task descriptions with What, Where, How, Why, and Verify elements.
berlysiareact-hooks
Guide React hooks refactoring to minimize effects and memoization.
berlysiasetup-claude-skills-for-web
Automate Claude Code skills setup from berlysia/dotfiles for web projects.
berlysiacodex-review-mcp
Automate external plan and architecture reviews via Codex MCP.
berlysiacodex-review-cli
Provide read-only code and architecture analysis via Codex CLI.
berlysiacommit-conventions
Analyze complex code changes and structure them into atomic conventional commits with clear scope and messages.
berlysiavalidate-plan
Validate blocked implementation plans using logic-validator and external reviews.
berlysiarenovate-github-action-guide
Configure Renovate with GitHub Actions using the official guide.
berlysiabugfix
Reproduce bugs, write failing tests, and verify minimal fixes via TDD.
berlysiadecompose
Decompose complex tasks into structured todos with completion criteria.
berlysialogic-validation
Validate decisions and claims by forking context for independent subagent review.
berlysiarefine-skill
Refine Claude Code Skills by rewriting metadata and workflows per CHECKLIST.md.
berlysiatask-handoff
Generate a shell command to replicate the active task list in a new Claude session.
berlysiaupdate-auto-approve
Analyze PermissionRequest decision logs to update auto-approve configurations across all layers.