From twinmind
Reference for deprecated TwinMind post-op pipeline skill, formerly an LLM subagent for shared post-operations. Now implemented programmatically via Node.js script tm-post-op.mjs with --layer (knowledge|action|both) and --event JSON options.
npx claudepluginhub volderlu/twinmind --plugin twinmindThis skill uses the workspace's default tool permissions.
Post-op pipeline 已程式化為 `.claude/twinmind/bin/tm-post-op.mjs`，不再使用 LLM subagent 執行。
Extends Claude Code via Postiz SKILL.md file. Uses frontmatter description '../../SKILL.md' for auto-activation matching in relevant conversations. Invoke directly with /Postiz. Useful for developers extending Claude Code capabilities in plugin directories.
Provides shared reference documents on CLI availability checks, fallbacks for bd/ao/gt/codex/cass, and multi-agent capabilities. Loaded JIT by AgentOps skills like council/crank/swarm.
Chains tools and AI models into bash pipelines for multi-step workflows like research-to-summary, code review-to-fix, conditional branching, and parallel multi-agent analysis.
Share bugs, ideas, or general feedback.
Post-op pipeline 已程式化為 .claude/twinmind/bin/tm-post-op.mjs，不再使用 LLM subagent 執行。
呼叫方式：
node .claude/twinmind/bin/tm-post-op.mjs --layer <knowledge|action|both> --event '<JSON>'
.claude/twinmind/bin/tm-post-op.mjs 是由 SessionStart hook 在 vault 端寫入的 shim，會 spawn plugin 內 scripts/post-op.mjs (pipeline 主程式，與 scripts/lib/ 目錄共構)。需要查實作時，從 .claude/twinmind/config.json 的 pluginRoot 欄位取得 plugin 安裝路徑後，再去讀 <pluginRoot>/scripts/post-op.mjs 與 <pluginRoot>/scripts/lib/。