Agent Advisor

Consult a cold-context advisor subagent — Fable or a fresh Opus — for a plan, a diagnosis, a critique, or a decision, while keeping Fable quota usage low. Use when the user says "ask fable", "ถาม fable", "ask opus", "ถาม opus", "consult an advisor", "ปรึกษา advisor", "agent-advisor", or asks for an architecture plan / debugging strategy on a genuinely hard, high-stakes, or hard-to-reverse problem. For reviewing or sanity-checking work that already exists, use scrutinize instead. Runs a gate first to decide whether the task needs an advisor at all (an explicit user request always passes the gat

10 stars 1 forksAI, LLMs & Agents Plugin marketplace

Install

Claude Code plugin
/plugin marketplace add claude-skills-kit/claude-skills-kit

The repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add claude-skills-kit/claude-skills-kit --skill agent-advisor
Manual install
git clone --depth 1 https://github.com/claude-skills-kit/claude-skills-kit.git /tmp/claude-skills-kit
mkdir -p ~/.claude/skills
cp -r /tmp/claude-skills-kit/agent-advisor/skills/agent-advisor ~/.claude/skills/agent-advisor

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Agent Advisor: cold-context second opinions, spent economically

A fresh `Agent` subagent sees only what you hand it. That is the whole
mechanism this skill sells: an advisor with no memory of this conversation,
no attachment to the approach already tried, and — at the top tier — more raw
capability than the main model.

Two advisor tiers:

- **Fable** — when the task needs *more capability than the main model has*.
  Fable is the strongest and most expensive model available; on a Max 20x plan
  its quota is roughly half the total budget, so it gets spent like a scarce
  resource: for genuine planning/debugging leverage, never for routine work.
- **A fresh Opus subagent** — when the task needs *a view the main model can't
  have*. The main model is often already Opus; when it is, an Opus subagent is
  **not** a capability upgrade and **not** a cheaper tier. Its only value is
  the cold context — no anchoring on this transcript, no sunk cost in an
  approach already tried — or standing in when Fable is unavailable. If
  neither applies, spawn nothing and think it through yourself.

**Not a replacement for the built-in `advisor()` tool** — a different
instrument. `advisor()` sees the whole transcript automatically, costs no
briefing, and offers no model choice; reach for it as a continuous
checkpoint ("am I on track, did I miss something"). This skill costs a brief
and lets you choose both what the advisor sees and which model answers; reach
Read the full skill on GitHub

Details

Skill name
agent-advisor
Path
agent-advisor/skills/agent-advisor/SKILL.md
Repo updated
29 September 2026

FAQ

What does the Agent Advisor skill do?

Consult a cold-context advisor subagent — Fable or a fresh Opus — for a plan, a diagnosis, a critique, or a decision, while keeping Fable quota usage low. Use when the user says "ask fable", "ถาม fable", "ask opus", "ถาม opus", "consult an advisor", "ปรึกษา advisor", "agent-advisor", or asks for an architecture plan / debugging strategy on a genuinely hard, high-stakes, or hard-to-reverse problem. For reviewing or sanity-checking work that already exists, use scrutinize instead. Runs a gate first to decide whether the task needs an advisor at all (an explicit user request always passes the gat

How do I install the Agent Advisor skill?

Claude Code plugin: /plugin marketplace add claude-skills-kit/claude-skills-kit. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Agent Advisor popular?

It lives in claude-skills-kit/claude-skills-kit, which has 10 GitHub stars and 1 forks.

Related skills

Agentforce Implementation Advisor

24

forcedotcom/partner-ecosystem-agentic-kit

Use for Agentforce agent, topic, action, prompt template, orchestration, state, testing, packaging, scratch-shape, and UX implementation decisions. Owns Agent, topic, action, orchestration, state, testing, and UX implementation decisions.

AI, LLMs & Agents

Wp Spec to Goal

23

nathanonn/agent-skills

Convert a WordPress plugin or feature idea (even a vague one) into a Codex /goal-ready bundle — GOAL.md, VERIFY.md, PROGRESS.md inside goals/<slug>/ — with wp-env + playwright-cli + wp-eval verification baked in. Asks clarifying questions in focused batches (each with options + recommendation + reasoning), reaches 95% confidence, optionally scaffolds a missing plugin folder + .wp-env.json + package.json + AGENTS.md, and emits a tailored /goal command. For new WP plugin or feature work, not bug fixes or non-WordPress projects.

AI, LLMs & Agents

Headless MCP Advisor

24

forcedotcom/partner-ecosystem-agentic-kit

Advise Salesforce partners on Headless Toolkit and MCP client/server choices, managed-package boundaries, HXL output, and delivery to Agentforce, Slackbot, Claude, or ChatGPT. Use for architecture decisions or Salesforce DX assessments; route builds to global sf-skills.

AI, LLMs & Agents

Moufa Xingyun Avatar

11

lewislulu/moufa-xingyun-avatar

给任意 Agent 接上一个魔珐星云(Xmov)实时 3D 数字人前端——网页里 new 一个数字人,把 Agent 的文字输出喂给 speak(),它就实时说话、配口型、做动作。当用户想"给 Agent / 聊天机器人 / LLM 加一个会说话的虚拟形象、数字人、虚拟主播、3D avatar 前台",或要接入魔珐星云 / 星云 / Xmov / XmovAvatar SDK 时使用本技能。

AI, LLMs & Agents

Implementation Best Practices

24

forcedotcom/partner-ecosystem-agentic-kit

Route and synthesize implementation decisions for Agentforce, Data 360, Slack, Tableau Next, Apex, APIs, prompts, MCP, sub-agents, state, long-running actions, testing, and user experience. Use after product and environment decisions are known and a partner needs a scalable implementation pattern.

AI, LLMs & Agents

Triton Kernels

45

j4flmao/agent-skills

Use when the user asks about OpenAI Triton kernels, @triton.jit, tl.load, tl.dot, triton.autotune, fused kernels in Python or torch.compile Inductor kernels. Produces a kernel, PyTorch wrapper, correctness test and do_bench benchmark. Not for Triton Inference Server (use ml-model-serving) or CUDA C++ (use cuda-programming).

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.