Kohei Okamoto
Community@Cohey0727
Kohei Okamoto's skill registry automates Git pull request lifecycles, multi-LLM code review, CI failure resolution, and Japanese freelance invoicing PDFs.
Agent Skills by Kohei Okamoto
Showing 21 vetted skills indexed across 1 GitHub repositories.
Cohey0727commit
Analyzes git changes and creates conventional commit messages with branch suffixes.
Cohey0727semantic-stacked-pr
Split one change into a stack of PRs by meaning using gh stack.
Cohey0727resolve-conflict
Detects, analyzes, and safely resolves git merge conflicts with user confirmation.
Cohey0727dev-issue-implement
Implements reviewed GitHub issues in git worktrees with tests and linked pull requests.
Cohey0727dev-issue-review
Reviews GitHub issues for validity, verifies claims against the codebase, and applies labels.
Cohey0727dev-hotl
Autonomously selects and executes one prioritized dev pipeline task from GitHub PRs and Issues.
Cohey0727resolve-ci
Diagnose and fix failing CI checks on GitHub pull requests using gh CLI.
Cohey0727git-workflow
Automates branch creation, commit, push, and GitHub pull request generation.
Cohey0727freelance-invoice
Generates Japanese qualified invoice PDFs from billing data using an HTML template and headless Chrome.
Cohey0727fanout
Splits files across parallel agents to apply one mechanical instruction tree-wide.
Cohey0727dev-issue-create
Creates researched GitHub development issues via the gh CLI after codebase investigation.
Cohey0727resolve-reviews
Fetch GitHub PR review comments, apply code fixes, and reply or resolve threads via gh CLI.
Cohey0727google-calendar
Manage Google Calendar events, availability, and invitations through MCP tools.
Cohey0727deep-review
Runs five parallel perspective-based subagent code reviews on a git worktree and merges findings into one verdict.
Cohey0727dev-pr-review-resolve
Resolves GitHub pull request review comments, applies fixes, and merges the PR.
Cohey0727evidence-redteam
Validates claims by gathering confirming and disconfirming evidence, then red-teams the reasoning.
Cohey0727discussion
Orchestrates a three-round debate among multiple LLMs to converge on one conclusion.
Cohey0727fusion-review
Runs parallel code reviews across multiple LLMs and merges findings into one report.
Cohey0727dev-pr-review
Reviews open GitHub pull requests in a worktree and merges, fixes, labels, or closes them.
Cohey0727freelance-work-record
Generates monthly work-record PDFs in a vertical calendar layout from daily time entries.
Cohey0727resolve-pr
Orchestrates resolution of CI failures and review comments on GitHub pull requests.
Frequently Asked Questions About Kohei Okamoto
FAQPage SchemaWhat tasks can I automate using Kohei Okamoto's skills?▼
You can automate the full Git PR lifecycle: conventional-commit message generation, stacked PR splitting, merge conflict resolution, CI failure diagnosis and fixes, review comment responses, Issue creation and implementation in worktrees, plus parallel multi-LLM code reviews and freelance invoice/work-record PDF generation.
Who are these skills designed for?▼
Software developers and freelance engineers working in GitHub-centric teams, especially Japanese-speaking developers. The dev-* pipeline skills target engineers managing Issue-to-PR flows, while freelance-invoice and freelance-work-record serve independent contractors needing Japanese qualified invoices.
How do the autonomous development pipeline skills work in practice?▼
The dev-hotl skill runs the dev pipeline (PR review, review response, Issue implementation, Issue review) as one autonomous task without human confirmation. Individual skills like dev-issue-implement use git worktrees to implement reviewed-labeled Issues, run tests, and open PRs.
What external dependencies do these skills require?▼
Skills rely on the GitHub CLI (gh, including gh stack for stacked PRs), git worktrees, and the Google Calendar MCP server for schedule management. Several skills invoke multiple LLMs in parallel for fusion-review, deep-review, and three-round discussion moderation.
Do these skills support multi-model code review?▼
Yes. fusion-review sends the same review request to multiple LLMs in parallel and merges results; deep-review launches perspective-specific sub-agents on worktrees; discussion runs a three-round multi-LLM debate with the session model acting as moderator to arbitrate and consolidate.