菜单
Manage Codex-native vault interactions without copying vault data or storing RED-sensitive material.
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 | vault |
| description | Manage Codex-native vault interactions without copying vault data or storing RED-sensitive material. |
Use the external Obsidian vault as durable memory while keeping this repository free of private vault data, secrets, and raw sensitive context.
Repo-local artifacts may include architecture notes, scorecards, templates, migration checkpoints, or sanitized examples. Runtime state belongs under ~/.ralph-codex. Durable human memory belongs in ~/Documents/Obsidian/MiVault.
Do not copy vault pages into this repo. Link or summarize only when the content is safe.
GREEN covers public facts, public docs, generic engineering patterns, or non-sensitive repo structure. YELLOW covers sanitized project context, redacted logs, or local decisions without secrets. RED covers secrets, API keys, credentials, private keys, wallet material, customer data, raw sensitive logs, or anything the user marks sensitive.
RED content is never externalized, never written to the vault, and never stored in repo artifacts.
Classify the content before saving anything. Keep RED local to the active Codex session. Write only sanitized GREEN or YELLOW summaries to ~/.ralph-codex or the vault. Handoff notes should name the source paths, decisions, validations, and unresolved risks without embedding private content.
Classification is content-aware. If a note requested as GREEN or YELLOW contains API keys, JWTs, private keys, seed phrases, wallet material, OAuth tokens, database URLs, .env references, or customer-sensitive markers, treat it as RED and skip persistence.
No secrets or raw vault data were written. Saved memory is classified and sanitized. The checkpoint explains the decision without needing private material.
Improve, 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.
