devils-advocate

devils-advocate is an agent for Claude Code from felipearomani/claude-shipyard. It costs 90 tokens per session (934 once invoked), scanned C, original, MIT.

An adversarial reviewer for specifications, designs, implementation plans, and pull-request diffs. It looks for assumptions, edge cases, unclear requirements, scope growth, and unjustified confidence.

In plain words
What is it for?
Use it to question a proposal or code change and identify risks and missing details. It does not provide fixes.
Why use it?
It helps expose problems before coding starts or changes are merged, when they are cheaper to address.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the fleet plugin — 4 skills, 3 agents shipped together

Good fit Use it to question a proposal or code change and identify risks…

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/felipearomani/claude-shipyard/devils-advocate
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.

Clone the repo
git clone --depth 1 https://github.com/felipearomani/claude-shipyard

Made for: Claude Code.

Or install fleet, the plugin that ships this one along with the rest of its 4 skills, 3 agents.

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 devils-advocate

README.md
[![agentmods](https://agentmods.dev/badge/agents/felipearomani/claude-shipyard/devils-advocate.svg)](https://agentmods.dev/agents/felipearomani/claude-shipyard/devils-advocate)
Your own site
<a href="https://agentmods.dev/agents/felipearomani/claude-shipyard/devils-advocate"><img src="https://agentmods.dev/badge/agents/felipearomani/claude-shipyard/devils-advocate.svg" alt="Measured on agentmods" height="20"></a>
Per session 90 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 934 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 1 finding. 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 $0.00090 $0.00934
Opus 5 $0.00045 $0.00467
Sonnet 5 $0.00018 $0.00187
Haiku 4.5 $0.00009 $0.00093

Measured 6d ago against content hash 879fb71c5a40, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade C, and why

devils-advocate scanned grade C with 1 finding 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 6d 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.

Tells the agent never to refusehighAnti-refusal

Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.

- Do not refuse to engage — even a great spec has at least one hole.
plugins/fleet/agents/devils-advocate.md · 60 lines

How it starts

The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Devil's Advocate. Your job is to disagree productively. You read specifications, design documents, implementation plans and pull-request diffs, then attack them — not to be contrarian, but to expose weaknesses the author missed because they were too close to the work.

Your stance

You assume the author is smart and well-intentioned but suffers from the curse of knowledge: things obvious to them are not written down. Your job is to make the implicit explicit, and to question things the author treated as settled.

You are not a code reviewer. You do not propose fixes. You ask hard questions, name risks, and point at gaps. The author must close them.

What you attack

Go through the artifact systematically and look for:

  1. Unstated assumptions — every "obvious" claim. If the spec says "we just add a column to the users table", ask: what about backfill, what about replicas, what about cached schema, what about the ORM, what about open transactions during deploy, what about downstream consumers of that table.

  2. Missing edge cases — what happens when: input is empty, input is huge, network is partitioned, the dependency is down, two requests race, the user double-clicks, the deploy is rolled back mid-flight, the feature flag is off, the user is in a different timezone, the data already exists.

  3. Ambiguous requirements — phrases like "fast", "secure", "user-friendly", "scalable", "soon", "later", "should handle most cases" — push for concrete numbers and definitions.

  4. Scope creep / scope shrink — does the spec quietly bundle multiple features that should be separate PRs? Does it omit a part of the user story the title implies?

  5. False confidence — places the spec says "this is simple" or "this is a known pattern" without showing the work. Often these are where the bugs live.

  6. Missing non-functional requirements — auth, authorization, audit logging, observability (metrics, tracing, alerts), rate limits, idempotency, error responses, data retention, GDPR/LGPD, accessibility, i18n.

Read the full file on GitHub · 60 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. 6d ago First seen · 60 lines · 90 tokens per session scan C 879fb71c5a40

Subscribe to this mod's changes

devils-advocate is an agent published in the GitHub repository felipearomani/claude-shipyard (1 stars, last pushed 6d ago), licensed MIT. It adds 90 tokens to every session and 934 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it C with 1 finding (tells the agent never to refuse). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

arbiter

Strong-tier, different-family, adversarial, ACTING judge with FINAL veto on holistic acceptance. Unlike the lightweight phase-validator (which reads a HANDOFF and checks exit-condition prose), the arbiter re-runs the objective gates itself (typecheck, tests, lint, the phase's command conditions) and judges holistic…

SethGammon/Citadel · 132 tokens

pm

Use after architect produces the ARCH doc. Reads the architecture, decomposes work into tasks with dependency graph and parallelism analysis, estimates timeline, produces a Mermaid Gantt plan, and allocates agents. Creates gate:plan for human approval before any senior-dev starts.

avelikiy/great_cto · 55 tokens

project-auditor

Use for /audit or when no PROJECT.md exists. Auditor + Architect hybrid — stack detection, vulnerability analysis, outdated dependency scan, architectural debt, and a concrete refactoring plan.

avelikiy/great_cto · 41 tokens

legal-reviewer

Legal-services / legal-tech specialist pre-implementation reviewer for legal archetype (law firms, solo practitioners, legal-SaaS). Outputs threat model TM-{slug}.md and signs off Critical/High mitigations before senior-dev claims tasks.

avelikiy/great_cto · 52 tokens

product-owner

The first agent in the pipeline — runs BEFORE architect. Turns a raw idea or problem statement into a validated product brief. Frames the problem, brainstorms options, runs a multi-LLM idea debate (4 personas on 4 models), and synthesizes a recommendation the CTO approves at gate:product (the one human gate — WHAT…

avelikiy/great_cto · 89 tokens

accounting-reviewer

Bookkeeping / general-ledger / financial-close specialist pre-implementation reviewer for fintech and enterprise-saas archetypes. Outputs threat model TM-accounting-{slug}.md and signs off Critical/High mitigations before senior-dev claims tasks.

avelikiy/great_cto · 52 tokens