council-of-rome

Council of Rome

Five Roman magistrates walk into your codebase.

A suite of opinionated AI evaluation skills that deploy five Roman magistrates to assess projects across architecture, UX, security, strategy, and operations. No compliment sandwiches. No Greek flattery. Just the standard of the Republic.

The Five Magistrates

Magistrate	Command	Historical Figure	Domain
The Censor	/censor	Cato the Elder	Code quality, architecture, structural integrity, technical debt
The Tribune	/tribune	Gaius Gracchus	UX, accessibility, user advocacy, equity, the people's voice
The Praetor	/praetor	Lucius Cornelius Sulla	Security, compliance, legal exposure, data privacy, dependencies
The Augur	/augur	Marcus Tullius Cicero	Market timing, competitive landscape, strategic risk, viability
The Legatus	/legatus	Pompey the Great	DevOps, deployment, monitoring, disaster recovery, operations

The Senate (/council) convenes all five simultaneously and surfaces where they disagree — because that's where the most valuable strategic tensions live.

Install

Claude Code (Plugin Marketplace)

# Add the marketplace
/plugin marketplace add therealatreides/council-of-rome

# Install the plugin
/plugin install council-of-rome@council-of-rome

Claude Code (Manual)

Clone this repo and copy the skill folders into your Claude Code skills directory:

git clone https://github.com/therealatreides/council-of-rome.git
cp -r council-of-rome/plugins/council-of-rome/skills/* ~/.claude/skills/

Claude.ai (Upload)

1. ZIP any individual skill folder (e.g., skills/censors-decree/)
2. Navigate to Customize > Skills
3. Click + and upload the ZIP

Manual / Any Agent

Each skill is a standalone SKILL.md file. Copy it into your agent's skills directory — the format follows the open Agent Skills specification.

Usage

Invoke any magistrate individually:

/censor                          # Structural integrity review (Cato)
/tribune                         # User advocacy review (Gracchus)
/praetor                         # Security & compliance review (Sulla)
/augur                           # Strategic risk assessment (Cicero)
/legatus                         # Operations & deployment review (Pompey)

Or convene the full Senate:

/council                         # All five magistrates + Senate hearing
/council --magistrates=censor,augur   # Select specific magistrates

Each magistrate produces a structured report with Latin severity classifications, findings organized by Vice → Erosion → Decree, and a final verdict.

How It Works

Each magistrate has:

• A historical persona with a specific mandate, temperament, and blind spots
• Five evaluation lenses tailored to their domain (e.g., the Censor's Five Pillars, the Tribune's Five Lenses)
• Agent-specific framing for each sub-evaluation
• A softness inspection calibrated to their domain — catching Greek flattery (Censor), patrician excuses (Tribune), defense attorney language (Praetor), court astrologer optimism (Augur), or garrison mentality (Legatus)
• Severity classifications with Latin nomenclature
• A structured report template inscribed as Markdown
• A verdict scale from catastrophic to acceptable

The Council layer adds a Senate Hearing that identifies contradictions between magistrates — speed vs. safety, user desire vs. structural integrity, market timing vs. readiness — and surfaces these tensions as the most actionable output.

Unique Mechanics

Magistrate	Unique Feature
Augur	Core Assumptions Register — identifies what must be true for the venture to succeed, then tests each assumption. Includes a mandatory Kill Question.
Legatus	Three Mandatory Scenarios — 3AM incident response, key person departure, and 10x scale stress test.
Tribune	Explicitly identifies who is served and who is abandoned — the section "who is abandoned" must never be empty.
Praetor	Dependency Dossier — treats every third-party library as a potential collaborator with the enemy.
Council	Senate Hearing — surfaces where magistrates contradict each other, which is where real strategic tradeoffs live.

Repository Structure