メニュー
ReAct reasoning — explicit reason-act-observe cycle for investigation and debugging. Use when you need to interleave reasoning with tool actions systematically.
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。
SOC 職業分類に基づく
Score way matching (embedding and BM25), analyze vocabulary, and validate frontmatter. Use when testing how well a way matches prompts, checking cosine similarity or BM25 scores, inspecting the embedding engine status, or validating way files.
Check how much context window remains in this session. Use when you want to know token budget, context usage, or how much room is left before compaction. Also use proactively when working on long tasks to gauge remaining capacity.
Cite governance controls and justifications when making recommendations about code quality, security, commits, or documentation. Use when justifying why a practice matters, when a user asks "why do we do this?", when reviewing code against standards, or when recommending practices that align with NIST, OWASP, ISO, SOC 2, CIS, or IEEE controls.
Ship current work through the branch → commit → push → PR → merge → cleanup flow. Picks up wherever you are in the cycle. Use when the user says "ship it", "land this", "merge this", or invokes /ship.
Review what changed in Claude Code upstream and compare against this project's commits and ADRs. Surfaces opportunities, gaps, and ADR status drift. Discovery conversation, not compliance dashboard.
Strategy selector for structured reasoning. Lists available thinking frameworks and invokes the right one. Use when facing complex decisions, trade-offs, or when stuck.
| name | think-react |
| description | ReAct reasoning — explicit reason-act-observe cycle for investigation and debugging. Use when you need to interleave reasoning with tool actions systematically. |
| allowed-tools | Read, Bash, Glob, Grep |
Before beginning, register this think session:
# Check for active session
if [[ -f /tmp/.claude-think-session ]]; then cat /tmp/.claude-think-session; fi
If another session is active, ask the user to finish or abandon it first. Otherwise:
echo "react" > /tmp/.claude-think-session
Read the strategy definition and follow its stages in order:
cat ~/.claude/hooks/ways/meta/think/strategies/react.md
This strategy is cyclic (max 8 iterations). Work through the reason-act-observe loop until you have enough evidence to synthesize a conclusion.
After synthesizing a conclusion, clean up so the think way can fire again:
rm -f /tmp/.claude-think-session /tmp/.claude-way-meta-think-*"${CLAUDE_SESSION_ID:+-$CLAUDE_SESSION_ID}" 2>/dev/null