メニュー
Analyzes current conversation for user corrections, patterns, and preferences, then proposes skill improvements. Use after sessions with corrections or when patterns emerge from successful approaches.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SOC 職業分類に基づく
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/materkey/cc-plugins --skill reflectコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Enter explore mode - a thinking partner for exploring ideas, investigating problems, and clarifying requirements before you have a plan. Use whenever the user wants to think something through, weigh approaches, or understand a codebase before changing it, even if they don't say "explore". When the picture is clear, it hands off to the planning plugin's /planning:make command to produce a single structured plan - it does NOT create OpenSpec-style proposal/tasks/design files.
Analyze a codebase for algorithmic complexity and performance hotspots, then propose or implement safe optimizations without breaking behavior. Use when asked to scan many files, find inefficient loops, nested iteration, repeated scans, costly rendering/recomputation, N+1 queries, avoidable O(n^2) or O(n) operations, or reduce complexity such as O(n^2) → O(n log n) / O(n), while preserving tests, APIs, outputs, and maintainability.
One-shot ralphex: continue an active plan or create one from context, then launch execution without questions. Use when user says 'ralphex auto', 'ralphex go', 'автоматически запусти ralphex', or wants plan+execution in one step.
| name | reflect |
| description | Analyzes current conversation for user corrections, patterns, and preferences, then proposes skill improvements. Use after sessions with corrections or when patterns emerge from successful approaches. |
| disable-model-invocation | true |
| allowed-tools | Read, Edit, Glob, Grep, AskUserQuestion |
Analyze the current conversation to find signals for improving existing skills.
This skill handles patch-level remediation only — fixing steps, criteria, format,
or behavior of a skill that already exists. If the problem requires a new skill,
a subagent, a hook, MCP integration, or a CLAUDE.md rule, use /reflect:reflect-architecture
instead.
Look for:
Ignore:
For each signal, check:
If the answer to either is "no", note it as an architecture signal and suggest
running /reflect:reflect-architecture after this skill completes. Do not attempt
to create new skills, hooks, or CLAUDE.md rules from here.
For each signal that maps to an existing skill, show:
If nothing found — say so and stop.
Ask the user:
Edit the target SKILL.md files directly with the approved changes.
After editing, remind the user: if an edited skill was already invoked in this session, it must be re-invoked for the changes to take effect.