review-council

review-council is a skill for Claude Code from indranildchandra/claude-local-starter. It costs 39 tokens per session (1,643 once invoked), scanned A, original, MIT.

A review process in which several independent expert perspectives examine an architecture or design, discuss disagreements, and reach a shared verdict with the developer.

In plain words
What is it for?
Use it for code, architecture, or design reviews, especially when a change affects several components or teams.
Why use it?
It reduces reliance on one reviewer and brings competing concerns into the same decision.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter. Also seen: mentions subagents.

Good fit Use it for code, architecture, or design reviews, especially when a change affects several components or teams.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/indranildchandra/claude-local-starter/review-council
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add indranildchandra/claude-local-starter --skill review-council
Clone the repo
git clone --depth 1 https://github.com/indranildchandra/claude-local-starter

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for review-council

README.md
[![agentmods](https://agentmods.dev/badge/skills/indranildchandra/claude-local-starter/review-council/github.svg)](https://agentmods.dev/skills/indranildchandra/claude-local-starter/review-council)
Your own site
<a href="https://agentmods.dev/skills/indranildchandra/claude-local-starter/review-council"><img src="https://agentmods.dev/badge/skills/indranildchandra/claude-local-starter/review-council/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for review-council

Your own site · 80×15
<a href="https://agentmods.dev/skills/indranildchandra/claude-local-starter/review-council"><img src="https://agentmods.dev/badge/skills/indranildchandra/claude-local-starter/review-council.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 39 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,643 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00039 $0.01643
Opus 5 $0.00019 $0.00822
Sonnet 5 $0.00008 $0.00329
Haiku 4.5 $0.00004 $0.00164

Measured 11d ago against content hash 1c45edf1fb58, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

review-council scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/review-council/SKILL.md · 172 lines

How it starts

The opening of the file, as written. The whole thing — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Review Council Protocol

Token budget — read this first

Load personas on demand, never all at once.

  • There are 20+ persona files. Loading all of them would consume the entire context window.
  • In Phase 2 you select 3–6 personas. Load only those files — nothing else.
  • Each persona subagent receives only: its own persona file + domain profile + scope brief. No other context.
  • Subagents run in isolation — they do not inherit the parent session's full context.
  • Persona outputs returned to the parent are compressed (max 300 words each).
  • If you find yourself loading more than 6 persona files, stop and reduce the council size.

Overview

The council reviews code, architecture, or design decisions using independent expert personas that debate, challenge each other, and converge on a verdict. The human engineer is an active participant, not a passive recipient.

Minimum council size: 3 personas. Council size by complexity:

  • simple (single component, low risk): 3 personas
  • medium (cross-cutting feature, moderate risk): 4–5 personas
  • complex (system-level, high risk, multiple teams): 5–6 personas

Always include at least one of: staff-engineer, cloud-cost-architect, appsec-architect — cost and security angles must be represented in every council.


Phase 0 — Scope + Early Human Input

  1. If /design-review was called with a scope argument, use it as-is.
  2. If no argument: run a quick repo scan (1 Haiku subagent, gitnexus ctx_search or file glob) to identify the highest-risk or most recently changed area. Propose this as scope.
  3. Present to human and ask:
    • "Does this scope look right, or do you want to focus elsewhere?"
    • "What concerns you most about this? Any known constraints (deadline, regulatory, performance budget)?"
    • "Any personas you want added or removed from the council?"
  4. Record answers verbatim as the Phase 0 Human Brief — this shapes everything downstream.
  5. If the user requests a persona not in standard-personas/: generate it using the schema in user-generated-personas/README.md, save it to user-generated-personas/<name>.md globally, then use it in this session.

Read the full file on GitHub · 172 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 172 lines · 39 tokens per session scan A 1c45edf1fb58

Subscribe to this mod's changes

review-council is a skill published in the GitHub repository indranildchandra/claude-local-starter (13 stars, last pushed 1mo ago), licensed MIT. It adds 39 tokens to every session and 1,643 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.