Automates the complete Issue-to-PR lifecycle: start working on Issues, create branches, implement changes, run quality checks, and manage PRs — all through a single workflow. Essential for any /rite: command, workflow questions, or when working with Issues, branches, commits, or PRs. Activates on "start issue", "create PR", "next steps", "workflow", "rite", "Issue作業", "ブランチ", "コミット規約", "PR作成", "作業開始", "ワークフロー", "次のステップ". Use for workflow state detection, phase transitions, and command suggestions.
From ritenpx claudepluginhub b16b1rd/cc-rite-workflow --plugin riteThis skill uses the workspace's default tool permissions.
references/coding-principles.mdreferences/common-principles.mdreferences/contextual-commits.mdreferences/phase-mapping.mdreferences/session-detection.mdreferences/work-memory-format.mdSearches, 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.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
This skill provides context for rite workflow operations.
When activated, this skill provides:
Workflow Awareness
Command Guidance
Best Practices
Coding Principles
Common Principles
Detect current state from:
{type}/issue-{number}-*
{type} values: feat, fix, docs, refactor, chore, style, teststyle is used for code style/formatting changes (no logic changes)| State | Suggestion |
|---|---|
| On main/develop, no Issue | /rite:issue:create or /rite:issue:list |
| On feature branch, no PR | /rite:pr:create after work |
| PR open, draft | /rite:pr:review then /rite:pr:ready |
| Long session (30+ minutes elapsed) | /rite:issue:update |
| Sprint with Todo Issues available | /rite:sprint:execute to run Issues sequentially |
| Sprint with multiple independent Issues | /rite:sprint:team-execute to run Issues in parallel with worktrees |
| Want to recall past decisions or context | /rite:issue:recall or /rite:issue:recall {scope} |
Key Principle: Always apply
question_self_check(see references/common-principles.md) before asking questions. Most questions can be avoided through context inference and using sensible defaults.
Ask immediately (do not defer) when:
If questions arise during work, record them in the work memory comment under "要確認事項" (Items to Confirm):
### 要確認事項
1. [ ] {confirmation_item_1}
2. [ ] {confirmation_item_2}
Target: Minimize questions through context inference and sensible defaults. Issue Start: 0-1 (score C/D only), Implementation: 0, PR Review: 0-1 (critical decisions only). Record non-blocking questions in work memory.
See references/common-principles.md for detailed frequency table by phase.
Automatically detect work state at session start and notify if interrupted work exists.
See references/session-detection.md for details.
{type}/issue-{number}-* pattern)See references/phase-mapping.md for phase list.
See references/work-memory-format.md for work memory format.
Avoid common AI coding failure patterns: surface assumptions, manage confusion, push back when warranted, enforce simplicity, maintain scope discipline, clean dead code, plan inline, and address all discovered issues.
See references/coding-principles.md for the full principle list and details.
Reduce excessive questions: self-check necessity, use defaults when available, infer from context.
See references/common-principles.md for details.
Before executing any /rite:* command, run the preflight guard. Resolve {plugin_root} per references/plugin-path-resolution.md.
bash {plugin_root}/hooks/preflight-check.sh --command-id "{current_command_id}" --cwd "$(pwd)"
Replace {current_command_id} with the slash command being executed (e.g., /rite:lint, /rite:pr:review).
If exit code is 1 (blocked), stop execution and display the preflight output. Do NOT proceed.
All gh commands that accept --body or --comment parameters MUST use safe patterns to avoid shell injection:
--body-file with mktemp for multi-line contentreferences/gh-cli-patterns.md for detailed safe patternsNever pass user-generated content directly via --body or --comment flags.
This skill works with:
/rite:* commands