菜单
Consult external LLMs (Gemini, OpenAI/Codex, Qwen) for second opinions, alternative plans, independent reviews, or delegated tasks. Use when a user asks for another model's perspective, wants to compare answers, or requests delegating a subtask to Gemini/Codex/Qwen.
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
Complete operational workflow for implementer agents (Codex, Gemini, etc.) making code changes and writing tests. Drives all work through atomic commits — each loop operates on the smallest complete, reviewable change. Defines the Code Change Loop, Test Writing Loop, Lint Gate, and Issue Filing process with circuit breakers, severity levels, and escalation rules. Requires `cortex git commit` for all commits. Includes bundled provider-aware review scripts that keep same-model shell-outs as the last resort, plus a fresh-context Codex fallback for code review and test audit. Use this skill when starting any implementation task.
Combined visual design and user journey review: color systems, typography, spacing, information density, visual hierarchy, workflow stages, emotional curve, friction points, and daily-use optimization. Use when evaluating how an interface looks, how it serves its user over time, and whether the design earns its space.
Use this skill whenever working inside any of Nick's repos (under ~/Developer/) or whenever a conversation touches Nick's durable cross-cutting work — Atlas Crew product strategy/brand/planning, the Unanet job, job search, interviews, tax amendment, foreclosure/tenant situation, strategic positioning. Maps Nick's three Basic Memory projects (products, work, main) and the per-repo dev trackers, and says which layer owns which kind of question. Consult before reasoning from stored memory summaries about repo architecture, product decisions, active priorities, or ongoing personal/career state — stored memory is lossy; the trackers and the Basic Memory projects are the real artifacts and should be consulted first.
User-triggered orchestrator that walks the user through running multiple sibling mapping/audit/doc skills against a shared scope, then assembles their outputs into one navigable artifact. Coaches rather than auto-runs — at each step the orchestrator presents the next sibling skill, recommended scope inheritance, and approval gate; the user invokes the sibling and confirms before moving on. Resolves the prior arch-analysis lookup once at suite start so every sibling skill inherits the same scope. Records each run's output path in a suite manifest. Produces a combined navigation HTML that links to each sibling's standalone HTML for one-stop sharing. Ships two named recipes — "release-driver onboarding" (for tooling that drives releases) and "deployed-system onboarding" (for systems that get deployed) — and supports custom plans. Trigger on "run our mapping suite," "do a complete onboarding," "build the combined report," "I need to onboard to system X," "give me the full picture of this codebase," or when the us
Use when completing tasks, implementing major features, or before merging to verify work meets requirements - dispatches code-reviewer subagent to review implementation against plan or requirements before proceeding
Use when executing implementation plans with independent tasks in the current session - dispatches fresh subagent for each task with code review between tasks, enabling fast iteration with quality gates
| name | multi-llm-consult |
| description | Consult external LLMs (Gemini, OpenAI/Codex, Qwen) for second opinions, alternative plans, independent reviews, or delegated tasks. Use when a user asks for another model's perspective, wants to compare answers, or requests delegating a subtask to Gemini/Codex/Qwen. |
| keywords | ["Codex","Gemini","Qwen","another LLM","compare models","cross-check","delegate to model","external model","second opinion"] |
Use a bundled script to query external LLM providers with a sanitized prompt and return a concise comparison.
settings.json under llm_providers.second-opinion, plan, review, delegate).Always run --help first:
python skills/multi-llm-consult/scripts/consult_llm.py --help
Example: second opinion
python skills/multi-llm-consult/scripts/consult_llm.py \
--provider gemini \
--purpose second-opinion \
--prompt "We plan to refactor module X. What risks or gaps do you see?"
Example: delegate a review
python skills/multi-llm-consult/scripts/consult_llm.py \
--provider qwen \
--purpose review \
--prompt-file /tmp/review_request.md \
--context-file /tmp/patch.diff
Example: plan check with Codex (OpenAI)
python skills/multi-llm-consult/scripts/consult_llm.py \
--provider codex \
--purpose plan \
--prompt "Draft a 5-step plan for implementing feature Y."
references/providers.md