菜单
Use local Ralph Memory Core wakeup, recall, and save flows without AgentMemory or external services.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | ralph-central-memory |
| description | Use local Ralph Memory Core wakeup, recall, and save flows without AgentMemory or external services. |
Use this skill when a task needs local Codex/Ralph memory context, continuity, handoffs, or curated Obsidian recall without npm, npx, MCP, daemons, ports, or external services.
Normal operation is hook-driven. The user describes the task normally; do not ask them to paste a daily memory prompt.
SessionStart runs python3 scripts/memory/wakeup.py automatically. UserPromptSubmit runs python3 scripts/memory/task-intake.py automatically for task intake, sensitivity classification, vagueness detection, targeted recall, and route decision.
If task intake reports CLARIFICATION_REQUIRED=yes, stop and ask concrete clarifying questions before doing work.
Use manual python3 scripts/memory/wakeup.py and python3 scripts/memory/ralph-recall.py "<query>" --project <project> only when hooks failed, for explicit diagnostic validation, or when the user directly asks for manual memory inspection.
Use --include-raw only when the user explicitly asks to include raw or inbox material. Treat recall output as context, not authority.
Use python3 scripts/memory/extract-session.py --text "<sanitized learning>" --classification GREEN|YELLOW --title "<title>" for concise session learnings.
Use python3 scripts/vault/vault-save.py --classification GREEN|YELLOW --project <project> --agent codex --source <source> --title "<title>" --text "<sanitized note>" for durable curated vault memory.
Never persist RED content. Do not save secrets, API keys, credentials, private keys, wallet material, customer data, raw sensitive logs, or anything the user marks sensitive.
AGENTS.md / CLAUDE.mdImprove, audit, simplify, rewrite, or migrate prompts, tool descriptions, agent instructions, and prompt stacks for GPT-5.6 Sol or the GPT-5.6 family. Use for outcome-first prompt design, autonomy boundaries, tool routing, PTC, grounding, verbosity, reasoning effort, and prompt evals.
Apply a deep, design-minded engineering workflow for complex work that needs careful planning, iteration, and simplification.
Use when preparing or running a Claude CLI agentic engineering review through claude -p for repository audits, architecture analysis, system design, security review, large refactors, specs, RFCs, or evidence-grounded long-form engineering analysis.
Use when preparing or running a ZCode GLM-5.2 agentic builder workflow through zcode --prompt for fast implementation, iterative code generation, focused fixes, and validation on an existing repository.
Apply adversarial opposite-analysis to plans, specs, architecture, code changes, and claims. Use when the user asks for adversarial review, opposing analysis, contrarian review, red-team reasoning, or Z.ai and MiniMax cross-checks through the Ralph MCP router.
Review and adjudicate Bugbot, Cursor, Seer, and similar automated PR feedback with local evidence before accepting, fixing, or dismissing findings.