AgentSkillsCN

council

以代理为先的项目文档编写。构建并维护专为AI代理设计的文档架构——包括路线图、设计文档、日志等。在用户说“/docs”、要求建立项目文档、希望更新文档，或在启动一个尚无文档结构的新项目时使用此功能。同样适用于用户想要规划新功能（创建设计文档），或回顾已完成的工作（更新日志/路线图）时使用此功能。

SKILL.md
--- frontmatter
name: council
description: >
  Consult an advisory council of three AI personas — Cato (skeptic), Ada (optimist), Marcus (pragmatist) —
  backed by different frontier LLM agents (Gemini, Claude, Codex). Each persona runs as a separate agent
  process with full repo context and returns independent feedback. Use when the user says "/council",
  asks for a second opinion, wants feedback on code changes, needs a premortem, wants to pressure-test
  a decision, or asks "what do you think about this approach?" Claude may also proactively suggest
  consulting the council before major architectural decisions, risky deploys, or ambiguous trade-offs
  (but should ask for user approval first).

Council

Consult your advisory council: three AI personas backed by different frontier models, each with full repo context.

Personas

NameArchetypeCore questionDefault provider
CatoSkeptical Strategist"What could go wrong?"gemini
AdaExpansive Optimist"What could this become?"claude
MarcusPragmatic Builder"What do we do next?"codex

For full persona details, see references/personas.md.

How to invoke

Run scripts/counsel.py via Bash from the current working directory:

bash
# Consult one persona
python3 <skill-dir>/scripts/counsel.py cato "Should we use Redis or Postgres for session storage?"

# Consult all three (runs in parallel)
python3 <skill-dir>/scripts/counsel.py all "Review the changes in the last commit — are we missing anything?"

# Override a persona's provider
python3 <skill-dir>/scripts/counsel.py ada "Is this API design too minimal?" --ada-provider codex

Replace <skill-dir> with the absolute path to this skill directory.

When to consult whom

  • •Cato alone: Security review, risk assessment, premortem, "is this safe to deploy?"
  • •Ada alone: Product direction, UX decisions, "is this feature compelling enough?"
  • •Marcus alone: Implementation tiebreakers, scope cuts, "what's the simplest path?"
  • •All three: Major architecture decisions, launch readiness, "should we do X or Y?"

Choosing the right moment

Proactively suggest consulting the council when:

  • •The user is about to make an irreversible architectural decision
  • •There are multiple valid approaches with non-obvious trade-offs
  • •A deployment feels risky or under-tested
  • •The user explicitly asks "what do you think?" about a high-stakes choice

Always ask the user before invoking — say something like: "This feels like a good moment to consult the council. Want me to ask Cato/Ada/Marcus/all for their take?"

Handling responses

  1. •Run the script and capture output
  2. •Show the full, unedited response from each persona to the user — do not summarize or paraphrase their words
  3. •After showing all responses, briefly note key agreements or tensions across personas
  4. •Ask the user how they want to proceed — the council advises, the user decides

Provider configuration

Each persona defaults to a different LLM provider for genuine model diversity. Override with flags if a provider is unavailable:

  • •--cato-provider claude|codex|gemini
  • •--ada-provider claude|codex|gemini
  • •--marcus-provider claude|codex|gemini

The script has a 5-minute timeout per persona. All agents run in read-only mode in the current directory.