Junya Morioka
Community@mjun0812 · Tokyo, Japan
Ph.D. in computer science.
Agent Skills by Junya Morioka
Showing 29 vetted skills indexed across 1 GitHub repositories.
deep-research
Researches a topic via web search and writes a cited Markdown memo in Japanese.
wezterm-control
Controls WezTerm panes, tabs, and windows via the wezterm cli command interface.
grill-self
Resolves planning and design decision branches through autonomous investigation and self-questioning.
experiment-plan
Creates machine learning experiment plans through structured interviews and saves them to .mjun/experiments/.
grill-me
Interviews users with sequential questions to resolve every decision point in plans and designs.
exmd
Generates single-file GFM explainer documents with glossary, TOC, and Mermaid diagrams from a fixed template.
exhtml
Generate self-contained HTML explainer pages with glossary, table of contents, and diagrams.
chat-note
Summarizes conversation findings into a single Japanese Markdown note file.
doc-sync
Detects and updates documentation that has drifted from the actual code implementation.
claude
Invokes the Claude Code CLI non-interactively for read-only consultation or delegated task execution.
codex
Invokes Codex CLI non-interactively for read-only consultation or delegated workspace-write tasks.
resume-other-agent
Reconstructs prior work context from another coding agent's session logs to continue interrupted tasks.
github-pr-create
Creates a GitHub Pull Request from the current branch with auto-generated title and description.
github-issue-polish
Refines GitHub issues into implementation-ready specifications through codebase investigation and trial implementation.
github-pr-review
Review GitHub pull requests with verified findings posted as reports and inline comments.
github-issue-create
Verify claims in draft material and create a labeled GitHub Issue from templates.
github-issue-resolve
Implements GitHub issues end-to-end from investigation through worktree-based development to pull request creation.
github-resolve-pr-comment
Fetch, classify, and resolve GitHub PR review comments with fixes and replies.
github-issue-update
Audits open GitHub issues and applies approved close, comment, and label changes in bulk.
github-fix-ci
Analyzes failing GitHub CI checks and applies fixes to affected files.
github-pr-fix
Detect and fix PR conflicts, CI failures, and review comments in a git worktree.
test-prune
Reviews all tests in a codebase and lists candidates for deletion or consolidation.
skill-review
Evaluates SKILL.md files against quality criteria and reports per-criterion verdicts with fixes.
git-commit
Review current changes and commit them with Conventional Commits formatted messages.
Frequently Asked Questions About Junya Morioka
FAQPage SchemaWhat tasks can I automate using mjun0812's skills?▼
You can create, polish, update, and resolve GitHub issues; create, review, and repair pull requests; squash commits and fix merge conflicts; run cited deep-research memos; generate Markdown or HTML explainers; sync docs with implementations; prune redundant tests; and proofread Japanese technical writing.
Who are these skills designed for?▼
Software engineers and researchers working in Japanese or English who drive coding agents through GitHub-centric development. The ML experiment-planning skill targets machine learning practitioners, while the writing skills target Japanese technical authors, bloggers, and book manuscript writers.
How do the GitHub and git skills operate at runtime?▼
They invoke the gh and git CLIs via scoped Bash permissions, plus jq, rg, and standard shell utilities. PR fixes run inside git worktrees, issue resolution chains implementation to PR creation, and bulk issue updates require user approval before applying label or close changes.
What prerequisites and dependencies do these skills require?▼
A git repository with GitHub CLI (gh) access for issue and PR skills; wezterm CLI for terminal pane control; Codex or Claude Code CLIs for cross-agent consultation and session resume; and python3, shellcheck, shfmt, or ruff for the skill-review quality checks.
Do the writing skills support new content creation or only editing?▼
Both. japanese-tech-writing and cognitive-rhythm-writing define style norms for drafting new Japanese technical chapters and articles, while stop-ai-slop-jp only rewrites existing AI-generated text, outputting a scored report and revision proposals without holding rules for fresh writing.