Skip to content
cccskills
FreePublished on GitHub

cursor-ask

Send a question, investigation, or design discussion to cursor-agent in read-only ask mode. It returns a second opinion or evidence-based findings that the host agent summarizes, without making file changes or running implementation work.

OriginalShow the original description

Read-only delegate to cursor-agent (Composer) for questions, investigation, design discussion, and adversarial sanity checks. No worktree, no cherry-pick, no Lead diff review — cursor-agent is locked to ask mode and cannot write. Use when user says: ask cursor, cursor sanity check, get a second opinion, adversarial review, design discussion, investigate with cursor, cursor:ask. Do NOT load for: implementation, refactor, file edits, commit/push work, anything requiring write access (use cursor:do or breezing --cursor instead).

Chachamaru127/claude-code-harness3,187 stars(GitHub 3,187 + cccskills 0)

Run score Not tested

See how to install

Good for

  • Ask for a second opinion on a design decision
  • Investigate code or configuration and request supporting evidence
  • Get an adversarial sanity check on a proposed change

Files included (1)

  • SKILL.md10.1 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

cursor:ask — Read-Only Cursor Delegate

cursor-agent (Composer) に read-only で質問・調査・設計相談・敵対的レビューを委譲する軽量スキル。

cursor-companion.sh task は引数なしで --mode ask (hard read-only stop) が自動で付くため、--write を渡さない限り cursor 側は ファイル書き込み・コマンド実行ができない。これにより worktree 隔離・cherry-pick・Lead diff review がすべて不要になる。

Quick Reference

cursor:ask "この設計判断、Composer 視点でどう思う?"
cursor:ask "TASK_BASE_REF からの diff を読んで、根拠のある見落としがあれば最大 3 つ挙げて"
cursor:ask "harness-mem の cross-project N-call、楽観的すぎる前提はある?"

用途:

ケース例
質問"この型エラーの根本原因は?"
調査"scripts/ 配下で curl を使ってる箇所を全部挙げて理由付きで"
設計相談"この abstraction、3 年後に保守できる?"
敵対的視点"この PR の最大の弱点を 1 つだけ挙げて"

Narration Rules (UX Contract)

敵は 冗長さ であって進捗報告ではない。起動時に何を聞くか・どう進めるかを簡潔に明示してから実行する。冗長な繰り返し・中身のない前置きだけを禁ずる。

起動時に必ず出すもの (banner + plan、3 行以内)

🚀 cursor / composer-2.5-fast / ask
これから: <質問の要点> を composer に投げて、結果を 3-5 行で要約

banner 1 行 + 計画 1-2 行。1 秒以内に出し、即 Step 2 へ。

進捗報告は出してよい

  • 委譲開始の 1 行 (→ composer に問い合わせ中)
  • 判断に必要な経緯を 1 行で

禁止 (= 冗長さ)

  • 同じ事実の 2 回言い換え: cursor-companion の結果を後段で再説明しない
  • 中身のない前置き: 「使い方を確認します」だけの行など tool call で自明な宣言
  • 3 行以上の経緯振り返り: 必要なら 1 行に圧縮
  • 起動シーケンス中の ★ Insight ブロック: Insight は最終要約で 1 回のみ

違反例 (冗長):

× 「cursor に質問を投げる準備をします」→ bash → 「投げます」(中身のない前置き + 言い換え)
× 「ask モードは読み取り専用なので安全です」と再説明(既知事実の繰り返し)
× ★ Insight ──── まず cursor の状態を確認します: ...

正常例 (簡潔 + 計画明示):

🚀 cursor / composer-2.5-fast / ask
これから: 設計の弱点を composer に問い、結果を 3-5 行で要約

Execution Flow

Step 0: 起動時 banner + plan

上記 Narration Rules に従い、banner + 計画 (3 行以内) を出してから Step 1 へ。

Step 1: banner 確認

Step 0 で banner + 計画 (3 行以内) は出し切っているので、ここでは banner 行が出ていることを確認する。banner は:

🚀 cursor / composer-2.5-fast / ask

以降は委譲開始の 1 行ステータス等で進捗を見せてよい。冗長な繰り返しのみ避ける。

composer-2.5-fast は scripts/model-routing.sh --host cursor --role worker --field model で解決される値の代表表記。実際の resolved model は cursor-companion 側のログに出る。

Step 2: helper root 解決 + cursor-companion 直接実行

$ARGUMENTS を質問文として渡す。--write は絶対に付けない。scripts/cursor-companion.sh を相対パスで呼ぶと consumer repo の cwd 直下に見えず exit するため、CLAUDE_PLUGIN_ROOT / HARNESS_PLUGIN_ROOT を hooks.json と同じ valid_root パターンで解決する (Issue #193 §2): 質問の目的、対象、既知の根拠を質問本文に添え、結論の理由と検証可能な証拠を求める。内部の思考過程は求めない。

QUESTION="$ARGUMENTS"
if [ -z "$QUESTION" ]; then
  echo "ERROR: question required. Usage: cursor:ask \"<your question>\"" >&2
  exit 1
fi

bash -c '
  set -euo pipefail
  valid_root() {
    [ -n "${1:-}" ] && [ -f "$1/scripts/cursor-companion.sh" ] && { [ -f "$1/.claude-plugin/plugin.json" ] || [ -f "$1/.codex-plugin/plugin.json" ] || [ -f "$1/.cursor-plugin/plugin.json" ]; }
  }
  HARNESS_PLUGIN_ROOT="${HARNESS_PLUGIN_ROOT:-${CLAUDE_PLUGIN_ROOT:-}}"
  ROOT="$HARNESS_PLUGIN_ROOT"
  if ! valid_root "$ROOT"; then
    ROOT=""
    if [ -n "${CLAUDE_SKILL_DIR:-}" ]; then
      probe="$(cd "${CLAUDE_SKILL_DIR}" && pwd)"
      while [ "$probe" != "/" ] && ! valid_root "$probe"; do
        probe="$(cd "$probe/.." && pwd)"
      done
      valid_root "$probe" && ROOT="$probe"
    fi
  fi
  if ! valid_root "$ROOT"; then
    ROOT=""
    for c in "${CLAUDE_PROJECT_DIR:-}" "$PWD" \
             "$HOME/.claude/plugins/marketplaces/claude-code-harness-marketplace" \
             "$HOME/.claude/plugins/cache/claude-code-harness-marketplace/claude-code-harness/"*; do
      if valid_root "$c"; then ROOT="$c"; break; fi
    done
  fi
  if ! valid_root "$ROOT"; then
    echo "ERROR: claude-code-harness plugin root not found (no scripts/cursor-companion.sh)" >&2
    exit 2
  fi
  HARNESS_PLUGIN_ROOT="$ROOT"
  bash "${HARNESS_PLUGIN_ROOT}/scripts/cursor-companion.sh" task "$1"
' _ "$QUESTION"

これだけで cursor-agent 側は --mode ask (hard read-only stop) に locked される。--force / --yolo も付かない。

Step 3: 結果を host が 3-5 行で要約

cursor の出力をそのまま貼らない。host (Claude/Codex) が読んで 3-5 行に要約 する:

  • 結論
  • なぜそう言えるか (cursor が挙げた根拠の核)
  • 注意点 / 追加調査が必要な点
  • 次の一手 (もしあれば)

最終回答は実際の結論、参照箇所、未確認点を返す。「これから要約する」という予告だけで終了しない。 助言は現物と照合して採否を判断し、相談だけの依頼から実装を開始しない。

Trust Boundary

cursor は不透明なサブプロセスであり、Harness のガードレール (R01-R13) は内部に適用されない。read-only 委譲でも以下の前提条件を満たすこと。

必須前提

項目内容設定場所
Secret 遮断.cursorignore で .env / *.pem / *.key / .ssh / .aws / .git を読取対象から除外repo root .cursorignore
Egress allowlist~/.claude/settings.json の sandbox.network.allowedDomains に *.cursor.sh を追加user settings
Filesystem allowlist同 sandbox.filesystem.allowWrite に ~/.cursor を追加 (cursor-agent が状態書込を行うため)user settings
permissions.json~/.cursor/permissions.json の terminalAllowlist / mcpAllowlist は read mode でも有効 (allowlist は best-effort、security boundary ではない)user config

詳細は .claude/rules/cursor-cli-only.md を参照。

ask mode で省略できるもの

通常の cursor 委譲で必要ask mode では不要理由
隔離 worktree不要cursor は書き込みできない
Lead diff review不要差分が生まれない
cherry-pick不要同上
worker-report.v1 / self_review 5 件不要実装をしないため

それでも残るリスク

  • 読み取り漏洩: .cursorignore を怠ると秘密ファイルが cursor 推論に渡る
  • 誤った情報の鵜呑み: cursor 出力は untrusted。Step 3 の要約で必ず host が判断軸を残す
  • allowlist 過信: Cursor 公式は "Allowlists are best-effort convenience. They are not a security guarantee." と明言。allowlist に依存しない

Topology

Lead (Claude/Codex) ──[cursor-companion.sh task]──> cursor-agent (--mode ask, locked read-only)
       │
       └──[Step 3: 3-5 行要約]──> User

Worker 介在なし。Reviewer 介在なし。worker-report.v1 / review-result.v1 契約は発生しない。

Related Skills / Rules

  • cursor-do — 書込タスク委譲(worktree + Lead review + cherry-pick の full containment)
  • breezing --cursor — Reviewer のみ cursor に逃がす lean second-opinion レーン
  • harness-review --cursor — レビューを cursor (composer-2.5-fast) に second-opinion として依頼
  • .claude/rules/cursor-cli-only.md — Cursor backend governance (trust boundary, prohibited flags)

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

agent-browser

FreeSummary

Use the agent-browser CLI to navigate web pages, inspect interactive elements, fill forms, capture screenshots, and collect page information. It helps check web app behavior through browser actions and snapshots.

  • Navigate a page and inspect its current state
  • Fill and submit a web form
  • Capture a screenshot for a UI check
Chachamaru127/claude-code-harness3.2KUpdated Oct 26, 2026

breezing

FreeSummary

Run planned software tasks as a coordinated team workflow. It assigns ready tasks to isolated workers, performs per-task and integrated reviews, incorporates approved changes, and reports progress and completion. Worker execution can use the selected backend, including optional Cursor delegation.

  • Run all remaining planned tasks as a team
  • Delegate a selected task range to workers
  • Use Cursor for worker implementation
Chachamaru127/claude-code-harness3.2KUpdated Oct 26, 2026

breezing

FreeSummary

Run planned work through a team workflow: delegate implementation to workers, coordinate independent review and revisions, then report the outcome. Select a task range or backend, or run all unfinished tasks in one command.

  • Run all unfinished planned tasks
  • Delegate a selected task range
  • Request implementation through the Codex or Cursor backend
Chachamaru127/claude-code-harness3.2KUpdated Oct 26, 2026

cc-update-review

FreeSummary

Review changes that integrate Claude Code or Codex updates. Classify each new Feature Table entry as implemented, inherited upstream, or tracked in Plans, and flag entries that are only documented without supporting work.

  • Review a Claude Code or Codex update integration
  • Check new Feature Table entries against implementation
  • Confirm future work is recorded in Plans
Chachamaru127/claude-code-harness3.2KUpdated Oct 26, 2026

ci

FreeSummary

Diagnose failed CI runs by reviewing logs, reproduction details, specifications, and recent commits. Identify whether the cause lies in implementation, tests, or the environment, then make authorized fixes without hiding failures by skipping checks.

  • Find the cause of a failed CI run
  • Trace a build error to a recent commit
  • Fix implementation issues behind failing tests
Chachamaru127/claude-code-harness3.2KUpdated Oct 26, 2026

cursor-do

FreeSummary

Delegate one implementation or file editing task to Cursor Composer in a dedicated Git worktree. Review its changes, then cherry-pick the reviewed commit into the current branch and update the matching plan entry when present.

  • Implement one feature with Cursor Composer
  • Refactor a file through Cursor
  • Delegate a specific code change
Chachamaru127/claude-code-harness3.2KUpdated Oct 26, 2026

See all skills by Chachamaru127

Report a problem with this skill