codex-ralph-vault-loop

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.

Find real bugs in local code, branch diffs, or named surfaces with evidence-first triage, regression-test guidance, and Codex-native verification.

Inspect and safely maintain local Codex App/CLI state with report-first, backup-first, archive-only workflows for sessions, logs, worktrees, config project entries, and handoff reminders.

Author a Lottie (Bodymovin) JSON animation that renders in a local skia player. Use whenever the user asks to create, generate, edit, or fix a Lottie animation, or asks for "an animation" to load.

Scout audit sweep migration recurring chore vague goal hook skill decision-log AutoResearch opportunities; propose Ralph-native tool paths before inline work.

Refine rough Codex /goal prompts into concise, verifiable execution contracts. Use for $goal-refiner, goal prompt improvement, or long-running work shaping. Enforce direct /goal payloads under 4000 chars; use GOAL.md/INPUT.md/STATE.md for larger work. Do not execute without separate approval.

Capture sanitized decisions and lessons into the local Obsidian vault.

Use when a difficult debugging, architecture, migration, or hypothesis-validation problem needs an external ChatGPT Pro or Oracle CLI second opinion after local inspection, always with dry-run first, minimal file selection, local safety scan, and explicit user approval before any real external run.

Guide Codex when an existing application needs Telegram Bot API integration without making Codex or this plugin the runtime.

Plan supervised multi-agent workflows with approval gates, packet ownership, integration, verification, and reusable local artifacts.

Run structured code/security review for local, branch, or commit diffs with explicit RED-safe routing.

Run structured code/security review for local, branch, or commit diffs with explicit RED-safe routing.

Create a durable repo-local handoff document with a portable prompt for a fresh agent.

Run a severe code review board that fuses structural code-quality review with real-user bug-risk analysis, evidence standards, P0/P1/P2 triage, and ship-readiness verdicts.

Remove AI writing patterns from prose. Use this skill when writing, drafting, editing, reviewing, or revising any text to eliminate predictable AI tells, slop, and formulaic patterns. Trigger this skill whenever the user asks to "deslop", "de-AI", "make it sound human," "remove AI patterns," "remove AI tropes," "clean up AI writing," fix "slop," "deslop" text, or review prose for authenticity. Also use when the user asks you to write or draft anything and wants it to sound natural rather than AI-generated. Common use cases include scientific writing (manuscripts, abstracts, cover letters, grant narratives, discussion sections, peer review responses), blog posts, newsletters, memos, reports, and any other substantial prose.

Create or adjust Playwright TypeScript E2E tests on macOS; run headed Chrome/Electron with QA pacing, QuickTime recording, videos, traces, and reports.

Choose the best safe MCP lane across Codex, Z.ai, MiniMax, official MCPs, and local tools using intent, sensitivity, verification value, and then cost.

Route sanitized work to intent-appropriate Z.ai, MiniMax, ralph_coding_models, official MCPs, and local Codex tools while keeping Codex main as final owner.

Coordinate Codex main across subagents, MCP tools, vault memory, gates, evals, and handoffs.

Use local Ralph Memory Core wakeup, recall, and save flows without AgentMemory or external services.

Create an evidence-grounded framing document from transcripts, call notes, stakeholder notes, or conversation records.

Turn a shaped project kickoff transcript into a builder-facing reference document organized by system territory.

Review existing requirements or convert raw context into high-quality requirements for PRDs, specs, user stories, acceptance criteria, BRDs, FRDs, requirement catalogues, and stakeholder notes.

Generate beautiful, self-contained HTML pages that visually explain systems, code changes, plans, and data. Use when the user asks for a diagram, architecture overview, diff review, plan review, project recap, comparison table, or any visual explanation of technical concepts. Also use proactively when you are about to render a complex ASCII table (4+ rows or 3+ columns) — present it as a styled HTML page instead.

Consolidate Ralph/Codex handoffs and ledgers into reviewable memory candidates.

Prepare complex Codex objectives before native /goal execution. Use this skill when the user asks to clarify, validate, de-risk, board, or autonomously pursue a broad Goal, objective, plan, audit, recovery, or long-running coding outcome. Do not use it for simple native /goal set, status, pause, resume, complete, clear, or budget requests unless the request is ambiguous, risky, multi-phase, or needs pre-execution intake.

Run Codex-native AutoResearch loops with versioned scorecards, durable session files, METRIC packets, ASI logging, and keep/discard/crash/checks_failed decisions.

Use this skill when the user asks Codex to design, redesign, prototype, implement, or visually improve a frontend/full-stack product, landing page, dashboard, app flow, presentation, pitch deck, one-pager, microsite, or UI from PDFs, PPTX decks, images, screenshots, Figma links, web references, brand assets, or an existing codebase. The skill must ask the necessary intake questions, inspect the current repository, extract a reusable visual system, propose a plan, build inside the current project, validate visually, and deliver a handoff. Do not use for purely backend tasks with no UI, design, document, or presentation output.

Run correctness, quality, security, consistency, and AI-output gates for Codex-native migration work.

Manage Codex-native vault interactions without copying vault data or storing RED-sensitive material.

Turn a MiVault spec note into a gated implementation plan and handoff.

Evaluate Codex workflow outputs with repository scorecards and hard gates.

Maintain and apply RASS v1 scorecards with fixed weights, hard gates, and anti-gaming protections.

Review completed Codex work for checkpoint readiness, memory hygiene, and reusable sanitized learnings.

Run session-start, active-session, and session-end memory handling for Codex-native Ralph workflows.