メニュー
Systematic knowledge extraction from open-source projects. Use when: user shares a repo/link to study, says 偷师/steal/学习/研究, or wants to analyze another project's patterns for adoption.
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
Evidence-gated completion check. Use before committing, creating PRs, or claiming work is done. Pairs with verification-spec at task start.
Launch a real interactive claude CLI session in a target directory (typically a worktree) with a pre-seeded prompt. Use when the current session cannot absorb the task context — e.g., N parallel worktree jobs that would blow the context window, or when you need the user to watch progress in a visible terminal. Each spawned session runs independently in its own Windows Terminal tab.
FileNotFound recovery protocol. Before reporting a missing-file error, suggest near-misses by basename similarity (difflib threshold 0.4) so typos and stale paths fail loudly with actionable hints.
Paste-by-reference guard for Edit calls. When new_string would echo >30 lines of existing file content, use {{file:path:start:end}} reference instead and expand before submission.
File-IO protocol for parent ↔ subagent communication. Parent writes input.txt + context.json, subagent appends to output.txt with [ROUND END] sentinels. Used with Monitor tool for live observation.
Memory write gate enforcing four axioms (Action-Verified / Sanctity / No Volatile / Minimum Pointer). Triggered before any write to SOUL/, .claude/memory/, or memory storage.
| name | steal |
| description | Systematic knowledge extraction from open-source projects. Use when: user shares a repo/link to study, says 偷师/steal/学习/研究, or wants to analyze another project's patterns for adoption. |
| origin | Orchestrator — earned through direct practice (see commit history) |
| source_version | 2026-04-18 |
You are conducting a steal (偷师) operation: extracting transferable patterns from external projects to strengthen Orchestrator. This is NOT evaluation ("should we use it?") — it's extraction ("what can we learn and adapt?").
Core mindset (from 39 rounds of practice):
Triviality Filter — If input is ≤ 3 words with no question/code/task, respond directly. Skip full protocol. Full spec:
SOUL/public/prompts/triviality_filter.md
@skills/steal/sections/01-preflight.md @skills/steal/sections/02-deep-dive.md @skills/steal/sections/03-extraction.md @skills/steal/sections/04-output.md @skills/steal/sections/05-index-update.md
